From: "George Stewart" gdstewart@worldnet.att.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 35mm vs Medium Format
Date: 10 Dec 1998
I once read (in an astrophotgraphy book by Walis & Provin) that a fine grain
4x5 negative is capable of holds 2,200
gigabytes of information. As
painting has not gone away with the advent of photography, and largeformat
still lingers with the advent of smaller formats, so will conventional
photography when digital expands to take a majority of the photo market
share. I think we shall see the demise of one-hour labs, and many camera
stores, as the distinction between computers and cameras (what will
eventually be though of as peripheral equipment), begin to blur, and we go
down to the local electronics store to purchase our new equipment.
Well any way, 2,200 Gig is a lot of RAM, and would take a very long time
just to load. Conventional photography is here to stay.
Date: Sat, 30 Jan 1999
From: Danny Grizzle danny@mogulhost.com
To: rmonagha@mail.smu.edu
Subject: Medium Format Negative Storage
Was checking out your medium format page. Very good thanks for all the
work.
One little omission, if you might have a suggestion.
After the photos are shot, what to do with the negatives?
More to the point, I just bought a Microtek ScanMaker 5 ($2500.00) with
direct glassless film scanning, but there is one unanticipated problem.
The film carrier for 2.25" films requires negatives/transparencies to be
cut into individual frames. So I need a new filing system, with
individual image sleeves. I don't know where to start. I also need some
kind of permanent labelling system so that I can correlate scanned image
names with the actual negative.
The negative carrier supplied is actually 6x9. No 6x7, 6x6, or 6x4.5
specialty negative trays are available.
Any help locating a film storage system would be appreciated. BTW - the
ScanMaker 5 quality surprised me! Very good, even compared to my Nikon
Super CoolScan used for 35mm.
Danny Grizzle
Date: Wed, 3 Feb 99
From: Godfrey DiGiorgi ramarren@bayarea.net
Subject: Re: [Rollei] 6x6 cm scan question
I've been unimpressed with the results from scanners in the $400 range,
but if you go up to the $800 range, the Agfa Duoscan T1200 does a pretty
decent job. I considered buying one recently but, while it's fine for
proofing and web scans, it isn't quite up to the quality level that I
want.
I've been studying scanners for 6x6 for several months now and I realize
I'm not going to be satisfied with less than a dedicated film scanner. So
I'm waiting until I can afford the Minolta Dimage Scan Multi (about
$2200). That's the least expensive dedicated film scanner for medium
format on the market. The better flatbed scanners with comparable
resolution and accuracy for doing negatives all cost in the same general
ballpark and up. The Agfa T2500 looks to be a delight, but around $4000
is a bit on the heavy side for me.
Godfrey
>I realize that MF scanners for good quality is $1K and up world,
>but I wonder what sort of scan quality I could get with a modest
>flatbed and transparancy adapter.
>
>My goals are fairly modest. I'd like to scan 6x6 cm negs to a PC
>to create an index sheet (12 images per 8.5x11" paper) and perhaps
>make 4x4" prints (4 images per 8.5x11 piece of paper). Is there
>something available in the $400 price class that will do this type
>of work or should I wait another year or two?
Date: Sun, 14 Mar 1999
From: lawrence lawrence@hoflink.com
To: panorama-l@sci.monash.edu.au
Subject: URL for Master spec/feature list and for most scanners
To the Pan group:
I came across this very effective site (Digital Eyes) that has a
comprehensive specification and feature page for most every
currently available scanner. It also points to the respective
manufacturer's website and offers user review comments. Very
helpful.
http://www.image-acquire.com/scanner/index.shtml
Lawrence
Date: Sun, 14 Mar 1999
From: CPeter1714@aol.com
To: panorama-l@sci.monash.edu.au
Subject: Scanning
I really hope anyone who talks about scanning has read Wayne Fulton's
book
called "A Few Scanning Tips" available at www.scantips.com.
I would say it is THE book on scanning with a flat bed scanner and
scanning
in general. 206 pages and I think it's about $15.00
Charles of Sioux City
From: jaffe@gte.net
Newsgroups: rec.photo.misc,rec.photo.digital
Subject: Re: Color calibration
Date: Fri, 22 Jan 1999
The ICC (International Color Consortium) standards for color calibration
are more and more widely supported. Apple pioneered their acceptance
with Colorsync, and now Microsoft supports their use under Win95, 98,
and NT with their ICM 2.0 module.
ICC profiles describe the color characteristics of cameras, monitors,
printers, and software. Photoshop 5 embraces ICC as its central
approach to color management. There are ICC tools specifically intended
to promote consistent, platform-independent color for browsers and
Websites.
For more info see:
http://www.color.com (International Color Consortium site)
http://www.microsoft.com/hwdev/devdes/icmwp.htm
http://www.apple.com/colorsync/
From: dwa652@aol.comnospam (DWA652)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: MF transperancy mounting?
Date: 9 Mar 1999
>Ive just begun shooting transperancy film in MF. What is the common
>practice regarding mounting the images. If I want to scan them at some
>later date will there be an advantage to storing them in negative sleeves?
>If I submit the image to stock agencies should they be in mounts? My
>nievity comes from a 35mm background. thanks.
I use cardboard (press and stick) BAIR mounts from The Stock Solution. Their
web page seems to be down right now (I keep a link to them under Photo
Manufacturer Links in the links section of my web page), but they can be
reached at 1-800-777-2076.
God Bless,
Don Allen
http://members.xoom.com/donallenfoto
From: Bob Wheeler bwheeler@echip.com
Newsgroups: rec.photo.equipment.large-format
Subject: Re: 6x7 cm vs 4x5" - doing a test, would appreciate suggestions
Date: Tue, 23 Mar 1999
This is a very difficult experiment to control,
and any differences you may see will be due to
things other than the format. Consider the
scanning alone.
An 8x10 print scanned at 175 l/i has 1750 dots on
the long edge, and thus the film has 1750/2.2=795
dpi, assuming the long side of a 6x7 film is 2.2".
Similarly a 4x5 film has 1750/4.9=357 dpi. Even if
you only scan at 1000 dpi, you will have more dots
in both cases than will be used by the printer,
and since the film resolution is considerably
greater than these, you will see no difference due
to format. My empirical studies indicate that
about 3 on film scan dots are needed to define one
line per millimeter, hence the 795 dpi translates
into 795/(3*25.4)=10 l/mm, and the 357 dpi to
375/(3*25.4)=5 l/mm. A 6x7 film that did not
deliver at least 60 l/mm would be a very poor film
indeed, as would a 6x5 film that did not deliver
at least 20 l/mm.
Even if you could control the many variables
involved, you will not begin to see differences
until the print size is increased considerably,
and even then you will run into problems with the
maximum capacity of scanners.
Bob Wheeler --- (Reply to: bwheeler@echip.com)
ECHIP, Inc.
Lars Finnstrom wrote:
> Hi!
>
> Does anyone know of a good test that has been made to show the differences
> between 6x7cm transparency film and 4x5 inch transparency film after it has
> been printed at let's say 175 lines/inch?
>
> What I want to find out is what the end result is, after it has been printed
> on a printing press (incl the whole chain of scanning, separations etc etc).
>
> Do you have any suggestions as to how to perform a test like this? Since my
> area of photography is architectural photography my idea is the following
>
> 1. Shoot a building on E100S 4x5 with my Arca Swiss F-line with a 150mm
> Rodenstock
> 2. Shoot the same building on E100S 6x7 with a 90mm Rodenstock
> 3. Have them scanned on a high end drum scanner in a similar way
> 4. Make chromalin test prints
> 5. Make real final print press examples - just like of you were printing a
> book
>
> (Use a print line/raster of 175 lines/inch, the most used for architectural
> high end publications)
>
> I want to use a building as example A and then do the same thing in the
> studio with some resolution charts, small details, textures etc.
>
> It would be interesting to see the differences - as printed in a book of
> around 8x10 inches size - between 6x7 and 4x5 on:
>
> 1/4 page
> 1/2 page
> 1/1 page
> Double page
>
> I have contacted a very quality oriented printing company with the very best
> equipment and skilled employees who are willing to take part in this
test if
> I decide to do it. They have decades of experience from printing art books,
> architectural books etc. Now I'm anxious to do it right and plan the test
> carefully before I go further. So I'd appreciate some help.
>
> There has been so much talk about 6x7 and 4x5 and of course 6x7cm has its
> practical advantages, no doubt, but I want to see how it REALLY turnes out
> at tthe end of the process. With the same subject, same film, same
> everything.
>
> Best regard
> Lars
Date: Fri, 25 Jun 1999
From: TLC tomclark@sprynet.com
Subject: [BRONICA] Medium Format Scanners for film
The Epson 636 carries a base price of $299 plus $100 for the transparency
adapter. It got rave reviews in Shutterbug magazine a couple of issues ago.
It will handle up to 4X5 in the transparency mode. Street price is better
than that.
Personally I have ordered the Microtek Scanmaker 4. I expect it to arrive
today perhaps. I tried one out at my friend's shop. It does a good job
for a
medium price scanner. Regularly priced at $700, it can be bought for $560
net after a $100 factory rebate if you order it from pcmall. Check it out at
www.pcmall.com Be sure to tell them that you saw this special price in
their catalog with priority code HX3855, otherwise they will try to quote
you the regular price. Pcmall's catalog number for this item is 44660. Offer
expires June 30th.
Unlike traditional flatbed scanners which place a lightbox on to of the
regular glass window, the Scanmaker 4 has a separate glass tray for
transparencies that slides inside the unit. Optical resolution is 600 X 1200
dpi. 36 bit input AND output. Optical density is 3.4. Scans legal size paper
for reflective originals. Maximum transparency size is a whopping 8X10.
Microtek also makes the Scanmaker 5 which has optical resolution of 1000 by
2000 dpi. Software interpolation of up to 8000 dpi. Pcmall's price on this
unit is $2269 and includes a free Iomega Zip 250 drive. Offer expires
6/30/99.
Tom Clark
Date: Thu, 24 Jun 1999
From: classicphoto@bellsouth.net
Subject: Re: [BRONICA] Digital Scanners
you wrote:
>Yeah, especially when it's your first roll of 6x6 Velvia. Beeeeautiful!
>
>Speaking of which, and this is a more general MF/digital issue, but does
>anyone know of a cheap and dirty way to scan medium format negs and
>slides? Is there any way to rig up a flatbed (reflective) scanner to do
>transparencies? I know there are specially-designed slide scanners out
>there, but they all seem to be EXTREMELY expensive. What does everyone do
>about getting slides to digital (skipping the expensive interneg printing
>route)? My local service bureau charges in blood for slide scans...
>Any ideas?
Rob, I just purchased a scanner based on a recommendation I got from
the Author of "Adobe Photoshop 5.0 for Photographers". I'm very pleased
with the quality. It scans prints, negatives, or transparencies up to
8x10. The dynamic range, which is equally as important if not more
so than resolution is 3.4. The scan quality is very near what I've
obtained from drum scans, and the cost is ridiculously low. 538.00
less a 100.00 rebate. (expires June 30 1999). The scanner is the
Microtek ScanMaker 4. it differs from most in that it has a
separate bay for film, and scans "glassless". All of the reviews
I've seen have been very positive.
Regards,
Tim
Classic Photography
Date: Fri, 02 Jul 1999
From: YDegroot@aol.com
To: panorama-l@sci.monash.edu.au
Subject: Re: Photo Printer
It used to cost 500 bucks, and everyone was raving about it.
The replacement for the EX is the 1200, it does bigger sizes (up to 12x44
inches) and makes better prints, at 6 picoliters per dot.
Yet, the EX is still an excellent model, and I believe takes up to 11 inch
wide paper, and can also do up to 44 inches.
At a little over 200, it's a steal.
Ed. note: thanks to Aaron for sharing this info!:
Date: Sat, 03 Jul 1999
From: Aaron Reece reece2@roanoke.infi.net
Subject: Scanner 36 bit 1200 dpi limitation
Hi, Robert. I did some checking on the Plustek Opticpro 9636T scanner,
which is a flatbed scanner with a built-in transparency adaptor, after
following the link to the manufacturer's site from your Medium-format
related Articles page ("Scanner 36 bit 1200 dpi"). Unfortunately, when
used to scan transparencies, the maximum scan size is 1x1.5 inches, or
slightly less than the dimensions of a 35mm slide, making it useless for
medium-format and larger applications. There is a similar model by Acer
called the 620PT (also available in SCSI(620ST) and USB(620UT) models)
which can scan an area 5x7 inches, and is slightly cheaper (around
US$150). Same basic specs, 600x1200 dpi max optical resolution, 36-bit,
density range of 3.0 (I believe). I have not tried it, nor have I seen
representative output from the transparency adaptor.
I found a review of the Plustek unit on the Ziff-Davis Computer Shopper
site ( http://www.zdnet.com/computershopper
) but unfortunately they
make massive use of frames, so it's impossible to link directly to a
particular review. Gotta keep you reading all those ads, I s'pose.
Cheers,
Aaron M. Reece
From Rollei Mailing List:
Date: Mon, 01 Nov 1999
From: Bob Shell bob@bobshell.com
Subject: Re: [Rollei] Off topic: film scanning and digital imaging
Bill,
You need to do some basic reading before diving into all of this.
Get a copy of _The Non-Designer's Scan and Print Book_ by Sandee Cohen
and Robin Williams, and _Real World Scanning and Halftones_ by David
Blatner and Steve Roth. I keep these two books near at hand in my
office.
For MF scanning you need either a high-end flatbed or a purpose-built
medium format scanner. I use the Linotype-Hell Saphir for medium and
large format scanning, but it is not worth a damn for 35mm. Minolta
just sent me their Dimage Scan Multi which looks like it would be the
trick for someone who only wanted to do 35mm and MF.
I don't know anything about the printer you mention. The photography
community has pretty much standardized on the Epson six-color printers
for top quality output. The archival inks for the six-color machines
will be out in a short while.
You may want to sign on to the Photo News Network's digital list to
get answers to the sort of questions you are asking.
Bob
....
From Rollei Mailing List:
Date: Thu, 16 Dec 1999
From: "imagineero ." imagineero@hotmail.com
Subject: [Rollei] (off list) OT Scanner
You're new at this whole newsgroup thing, aren't you john? you might have
made a new message thread with an appropriate subject line rather than
replying to the thread on 'flex intermittent shutter failure'. The 2
subjects are not really related. You have a little to learn about
ettiquete.
Are you shooting MF film primarily? The Dimage is an excellent scanner and
has gotten good wraps, but you might want to consider looking at one of the
new range of scanners from Microtek, which you will find at these addresses;
http://www.microtek.com/usi-sm4.html
http://www.microtek.com/usi-sm9600.html
http://www.microtek.com/usi-sm5.html
The scanmaker 5 is the top of the range at around $2,000 it has an optical
resolution of 1000x2000 which doesn't sound like a lot, but you'll find it's
plenty for medium format. The software interpolated resolution is
8000x8000. This scanner is unique because it scans film and prints, and can
scan a whole roll of 120 *at one time*. it can take film up to 8"x10".
The scanmaker 4 is a cheaper model at around $600 but you might find it
adequate for your needs. Either model can be mail order purchased at;
http://www.outpost.com
Search for 'microtek scanmaker'. I have had good experiences with this
mail order company. I think you'll find that the tax is pretty high to
import to the UK. Here in south Korea it's 30%, in Australia it's the
same. Sometimes your better off to buy locally, it's possible to get a
cash discount and you'll also get a warranty if anything goes wrong.
Imported items can also have power supply problems.
....
From Panoramic Mailing List:
Date: Wed, 08 Dec 1999
From: MIKE GRACE amazing50@hotmail.com
Subject: Archival Inks Source
You can purchase archival inks and papers for many printers from this
company. I found the link some time ago but haven't ordered anything form
them yet.
http://www.inkjetmall.com/store/index.html
Mike Grace
Date: Sat, 22 Jan 2000
From: Robert Erickson cirkut8@yahoo.com
To: panorama-l@sci.monash.edu.au
Subject: Re: Bulk Ink for Epson
Loading ink is easy, but can be messy. I recomend
wearing gloves.
http://www.inkjetrefills.com/BulkInks/epson.html
http://www.refillkits.com/EpsonBulkInks1.html
http://www.refills.com/usa/epson.html
http://www.netwares.com/bulk.html
http://www.jetink.com/
http://www.mrinkjetinc.com/epson1.htm
=====
Robert Erickson, cirkut8@yahoo.com
From Hasselblad Mailing List:
Date: Sat, 4 Mar 2000
From: DavidG6028@aol.com
To: hasselblad@kelvin.net
Subject: Re: Scanning Medium Format
david@scc.net writes:
I'm considering replacing
that with a new flatbed with a built-in transparency adapter. I don't have
much to spend though - maybe $600 max. Any models you'd recommend?
I just purchased a UMAX Astra 2200 (SCSI & USB) for my Mac. It was about
$180, and comes with an integrated attachment for scanning film up to 4x5.
It is NOT intended to replace a high end film scanner, but does a credible
job of scanning medium format media. I purchased it for my son so that
both of us could start a low cost introduction to digital image
processing. It also came with Photoshop LE. Very good value.
-David Gerhardt
Date: 27 Dec 1999
From: bhilton665@aol.com (BHilton665)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: medium format scanners and af cameras
>From: ira moore / marianne stolk ismo@hacom.nl
>
>i'm considering the move to medium format and was looking at the
>possibility of buying a medium format scanner (either the minolta mf or
>the nikon ls-4500)... does anyone have any experience with these
>scanners on a Macintosh system?
If you're talking about the Minolta Multi then just a note about the scan
resolution ... it's 2,820 dpi for 35 mm and APS but only 1,128 dpi for
medium
format.
This translates into the following pixel counts for 4 common formats ...
10.8 million pixels for 35 mm (2688 x 4032)
7.2 mpix for 6x7cm (even though the film area is ~ 4x larger than 35 mm!)
6.4 mpix for APS
4.6 mpix for 6x4.5cm (even though the film area is ~ 2.7x larger than 35
mm)
In other words you'll get 50% more pixels from 35 mm than from 6x7 cm. If
you
want to make prints at 300 dpi the 6x7cm won't even print an 8x10" full
frame
(2496x2880 pixels before the black edges are cropped out).
I think the Nikon has a rather low rez for MF and 4x5" film as well.
Getting high rez MF scans is a bit of an expensive bother right now,
unfortunately.
From Rollei Mailing List:
Date: Sun, 12 Mar 2000
From: Bob Shell bob@bobshell.com
Subject: Re: [Rollei] Re: Need a film scanner for my Rollei output
The reason scanner prices have stopped dropping and not many new ones
have come out is the disastrous earthquake in Taiwan where all of them
are made. Some factories were destroyed, some damaged, some key
personnel killed and injured, and many small subcontractors destroyed
or disrupted. This is all getting sorted out and by the fall I expect
to see some price movement and a lot of new scanners introduced.
There will be at least one medium format scanner with Digital ICE
technology then.
Bob
From Panoramic Mailing List:
Date: Wed, 29 Dec 1999
From: Glenn Barry glenn@acay.com.au
To: panorama-l@sci.monash.edu.au
Subject: Re: Printers - a list of those that can do long prints?
One more link that I just discovered that is very informative on the
subjest of film scanners and comparisons with an IT8/Q60 target.
http://www.cix.co.uk/~tsphoto/tech/filmscan/index.htm
Date: Mon, 08 Nov 1999
From: Dave Wyman mt.man@bigfoot.com
To: rmonagha@post.cis.smu.edu
Subject: Microtek Lightlid
Robert,
Have you seen the Lightlid, by Microtek? Replaces the lid of a few
flatbed scanners (X6, V300, V310, V600). Let's you scan transparencies
up to 5x6 inches...I just got one - it works wonderfully for putting
images on the web.
I'm going out of town tomorrow, or I'd spend the day experimenting -
unfortunately, I'll have to wait until next Sunday. (At least I'll be
away making photographs with my Rollei). I had time to try one image
made with my old Rollei - fantastic result! Cost: about $70, which means
I don't have to schlep over to my friend's place, to use his $2000
Minolta scanner.
http://www.microtek.com/lightlid.html
I love your web pages.
Dave Wyman
--
Image Quest Photography Tours
Family Adventure Tours
The Northern California Book Project
http://www.davewyman.com
From Panorama LIst:
Date: Fri, 24 Mar 2000
From: ralph fuerbringer rof@mac.com
To: panorama-l@sci.monash.edu.au
Subject: Re: MF Scanner Choices
best compromise deal now is epson 1200u (stands for unadorned dpi) lists
for 350 w/transparency adapter, does up to 45 film and includes a 6x9
carrie which they nostalgicallly call "browwnie size. Next step up to 1600
is 3 times the price 1/3 more dddp's.
-- rof
From panoramic mailing list;
Date: Tue, 13 Jul 1999
From: YDegroot@aol.com
To: panorama-l@sci.monash.edu.au
Subject: printing dpi revisited
This came froma website of a guy who shoots art nudes. I hope no one is
offended.
The PhotoShop tip site is really helpful.
FROM http://photoshoptips.i-us.com/Nudes.htm
Photoshop File Resolution
Keep in mind the basic guidelines of file resolution. For web pages, the
resolution needs to be on a 1:1 ratio with monitors, or 72 dpi. For grey
scale halftones, the rule of thumb is: the dpi (dots per inch,
equivalent to ppi or pixels per inch) should be twice the lpi (lines per
inch). So if you are preparing a scan for a newspaper that prints with
an 85 line screen, your file resolution should be about 170 dpi (at
final size). This rule generally holds true for color separations,
although most color houses know through experience that you don't need
more than about 1.75 times the lpi. For 150 lpi color separations, 265
dpi is quite adequate and anything larger is "wasted space." This is
because there are four sets of screens overlapping (intermeshing at
different angles) each other instead of just one set of dots for grey
scale images.
For a typical 300 dpi laser writer which prints only up to 53 lpi, 100
dpi grey scale scans are adequate. For the original Epson Stylus Photo
ink-jet printers (which image at 720 x 720 dpi), the maximum resolution
should be 240 dpi (and, according to Epson techs, anything higher is
re-sampled down for you by the printer's software). The newer Epson
Stylus Photo EX, imaging at 1440 dpi, recommends files up to 360 dpi for
optimum quality output. You may, however, be hardpressed to actually see
the difference between a 240 dpi and a 360 dpi file printed on an EX
unless your scan is a drum scan from a large format negative or
transparency. For continuous tone digital imaging, the necessary
resolution is relative to the printer's resolution.
[Ed. note: thanks to Aaron for sharing these tips!...]
Date: Tue, 13 Jul 1999
From: Aaron Reece reece2@roanoke.infi.net
To: Robert Monaghan rmonagha@post.cis.smu.edu
Subject: Re: Scanner 36 bit 1200 dpi limitation
Robert:
If you will be scanning strictly for web use, you might want to
consider the Microtek Scanmaker 4 (not 5). This is a 600 dpi, 36 bit
flatbed model that sports one of those under-the-glass transparency
trays like the Agfa top-of-the-line model. According to owners whose
opinions I've dug up on web sites, this works quite a bit better than
the built-in transparency adaptors on units like the Acer Prisa 620PT et
al. Price-wise, they're going for $500 on buy.com, which isn't too bad
if the output measures up.
Whatever you decide, let me know what you think as I am also
considering the purchase of a like unit.
Godspeed and many thanks for a very useful site.
-Aaron
postscript: I forgot to mention, the Scanmaker 4 allows scanning up to
8x10 inch transparencies.
-Aaron
From Contax Mailing List:
Date: Tue, 16 May 2000
From: "Bob Shell" bob@bobshell.com
Subject: Re: [CONTAX] OT: Scala
I haven't done the research, but David Brooks has. He says that
nearly all scanners have problems scanning silver images. Something
to do with the silver grains scattering the tightly focused light beam
of the scanner. The image in Scala is a silver image.
David recommends chromogenic black and white films (Ilford XP-2 Plus
or the Kodak chromogenics) if you plan to scan. It makes no
difference in scanning a negative or positive since you can switch
easily in the scanner software.
If you really want a positive image you can experiment with the
chromogenic films and having your lab run them in E-6. Scala sales
seem to be stagnant after an early peak because only a few places can
process it and the home processing kits are not readily available yet.
Bob
...
From Panoramic Mailing List:
Date: Wed, 31 May 2000
From: "M. Denis Hill" denis@area360.com
Subject: RE: Panoramic shots
For offset printing, you need to have your image scanned at "dpi" equal to
150% of the line screen of the printing. For high-quality color printing,
I assume that you'll print at 150 lpi (lines-per-inch) or more, so you
would have the image scanned at 225 dpi at the size it will be printed.
For printing, the file needs to be CYMK (four color), which has a
compressed color gamut compared to RGB. You can have the scanning service
perform the conversion (the native format of scanners is RGB), or do the
conversion in Photoshop. For truly high quality printing, you can go into
Hexachrome (six-color), but that's another ball of was. If you are a
novice at this, I recommend that the scanning service do both the CMYK
conversion and color correction. In any event, you should require a
matchprint from the offset negatives (from which printing plates are
made), to assure that What You See I What You Want.
Digital files can be made into transparencies in what's called an LVT or
film recorder. If you want one source for scanning and LVT output (after
you've done the stitching), you might try http://www.nancyscans.com/.
M. Denis Hill
Area 360 Communications
http://www.area360.com
From Contax Mailing List:
Date: Wed, 28 Jun 2000
From: "Bob Shell" bob@bobshell.com
Subject: Re: [CONTAX] Agfa Pan 100 - Good?
Be prepared for problems scanning regular black and white negs with
most scanners. The silver grains in the films scatter the focused
beam of the scanner. I get best results from scanning chromogenic
black and white films.
BTW, George Schaub has written a good book on black and white
digital imaging. I don't recall the exact title at the moment. He's
working on a sequel right now.
Bob
From: havriliak@aol.com (HAVRILIAK)
Newsgroups: sci.astro.amateur
Date: 05 Jul 2000
Subject: Re: CCD vs film resolution
> CCD pixels larger or
>smaller that film "Pixels"
The effective film "pixel" is called a grain, i.e. microcrystal of silver
halide, the basic unit of photography. In otherwords when a silver halide
grain is absorbed by a minimum number of photons, a developer acts on it
to reduce it to silver. The grain size is typically less than a micron.
A CCD pixel is of the order of 10 microns. The resolution of film is
superior to CCD's while the effective speed of a CCD pixel is an order of
magnitude faster than film
From Rollei Mailing List:
Date: Sun, 9 Jul 2000
From: CLARINETJK@aol.com
Subject: Re: [Rollei] Off-Topic...Scanning MF negatives?
Try www.epson.com. The Epson Expression 1600 comes in four models. The
upper
end two models (designated Pro) come with a transparency unit. Also, check
out June 2000 issue Shutterbug (on line?) for a review by David
Brooks. This
scanner is great for medium and large format negatives and transparencies.
However, I'm not going to throw out my 35mm film scanner. JK
From Leica Mailing List:
Date: Mon, 17 Jul 2000
From: Jim Brick jimbrick@photoaccess.com
Subject: [Leica] Re: Leica Users digest V17 #253
Fuji states, in their ad's, that it is 6mp "image size" which is quite
different from a 6mp sensor. Fuji uses a (I cannot remember off the top of
my head) 2.8mp sensor and generated the 6mp size by interpolation. Look up
"interpolation" in the dictionary.
Philips has had and it is used in many "pro" digital cameras, a true 6mp
sensor. We have one here (sensor that is) but six megapixels (divided by
four as all of these sensors are Bayer pattern) is still miles and miles
away from the capabilities of film. It will require a new sensor
technology to be able to compete with film. Current technology puts us at
the limit of the signal to noise ratio. We cannot make the current cell
components any smaller. The number of electrons stored to record a pixel
has to be enough to seen above the noise of semiconductor junctions and
capacitive layers. The process is already at the .10 micron level and is
beginning to conflict with the atomic structure of the molecules.
Therefore, to make a "higher resolution" electronic sensor, a new
technology will have to be developed. And it is being feverishly worked on
as we speak.
Keep buying those Leica lenses. It'll be a long time before film plays
second fiddle to digital.
Jim
LRZeitlin@aol.com wrote:
>I see from the trade press that Fuji has released a semipro digital SLR with
>a 6mb + pixel photosensor. Several months ago there was a thread on the LUG
>that suggested that a 6mb pixel density would be the threshold for film
>equivalent results in digital photography. Does anyone care to comment?
>
>LarryZ
From Leica Mailing List:
Date: Tue, 18 Jul 2000
From: Erwin Puts imxputs@knoware.nl
Subject: [Leica] digital prints again
Today I was in Solms and they showed me some prints from a very
professiona photographer who works on Velvia and K64. Prints were made
from scanned negatives and digitally printed with 152 lpmm and with an
intelligent procedure of rasterization. The quality is beyond what you can
get with Epson printers as a generic class. The print size was A4. They
looked beautiful, sharp, saturated colours etc, whatever you would like.
The eye ccould not ask for more and indeed, as I said in my previous post,
the limit of the eye's resolving power has been reached. Now I used my 10
x loupe and I did not see ANY detail, only raster points, and so did the
Leica people. I had with me some B&W; prints at 30x40 cm and when I used
the same loupe on tese images, any body saw detail, more information and
more detail into the detail. NO raster points or whatever, just plain real
detail.
I do agree with anybody on this list that a good digital A4 at normal
viewing distance will give the impression of exquisite detail, but it
simply is not there. The eye can not resolve it as this distance, that is
the limiting factor, If you need to see more detail, you have to enlarge,
which the film can handle and the digital print cannot. This level of
recording ability may not be of any interest to most observers of Leica
prints. To deny it is a different ball game.
I would indeed challenge anybody on the list to use a Leica negative, scan
it ar whatever resolution, print it digitally at whatever high end
industrial printer to a format of 30x40 cm and compare it to a chemical
print at the same size and look at it really close. Let alone go for a
slide show at a hundred times enlargement.
I agree that digital prints look convincing, and are in itself impressive.
I also find them wanting in detail at a level any chemical print can
exhibit.
I am not against digital prints and I indeed have a digital darkroom. When
you are used to look at fine detail and gradation at a 25 times
enlargement factor, the digital process is still far beyond the analogue
process.
My point is not that I am not willing to accept the claims of digital
excellence. I do. My point is that willing to express leica excellence is
still beyond the capability of digitally generated prints.
My challenge stands for the Boston LHSA meeting.
Erwin
From Leica Mailing List:
Date: Sat, 22 Jul 2000
From: Jim Brick jim@brick.org
Subject: [Leica] RE: Horology
To reflect on Erwins thoughts, I have two quick stories.
....[on digital vs analog watches joys..]
Besides being a photographer, professional from the standpoint that I went
to Brooks Institute of Photography, worked as a commercial photographer
for many years, currently teach workshops and take/make/and sell fine art
prints from 20x24 to 48x60, I am an engineer. I am in the middle of
designing and producing a digital camera for RCA. We have used all of the
latest digital sensors from 1/4" to 24mm square, from .5 megapixels to 6
megapixels.
Digital sensors cannot compete with film. I've given a dissertation on
this before so I won't repeat it now. But simply, and roughly, it takes
four pixels to record one true color pixel. Pixels cannot be made smaller
than 3 microns square. Four pixels = 36 square microns. So in 36 square
microns you get a single RGB 24 or 36 bit pixel.
The average size of a silver halide grain is one square micron. Some
larger, some smaller, but the average in normal film like APX 25 or
Kodachrome would be around one micron square. Within this one micron,
there are, on an average, 20 BILLION silver halide molecules that can be
struck by a photon and converted to silver. It takes several molecules
participating to record a "speck" of silver. So even if we divide by
twenty, we have a density range, within one square micron, of a billion.
Rather than 4096 levels of density (twelve bits per physical pixel) in
nine microns square or sixteen million combined color densities in 36
square microns.
The net result is that you cannot record fine detail using a digital
sensor unless you are willing to make multiple scanning passes, with a
micron or two shift of the sensor on each scan, and then use a very
sophisticated computer program to process the multiple images and ferret
out what the fine detail was. This is one of the techniques astronomers
use to capture fine detail in astrophotography. But alas. You need a
static subject, sophisticated camera sensor mechanics, and a very
sophisticated processing system.
Hardly an M6 with a 35 ASPH lens and Kodachrome or APX 25. Which will
still win the fine detail war, hands down. And you can carry it in your
pocket.
Jim
From: oorque@aol.com (OorQue)
Newsgroups: rec.photo.equipment.medium-format
Date: 22 Jul 2000
Subject: Re: Cost of move to MF
> Also when I thought about it, $40 per > scan could get me quite a few scans for
> the price of a $2000 scanner.
Don't forget about rentals. Here in Phoenix, there are at least two stores
that will rent a Minolta Multi-Scan for ~ $55/day and ~$85 for a weekend.
I realize this isn't the best scanner available but it should be be good
enough for 8x8 or even 10x10 prints -- Yes, I'm one of those 6x6 lovers!
-- and while it's no fun to spend hours on end scanning, it's a
cost-effective way to bide your time until an affordable, high-resolution,
MF-capable film scanner hits the market.
JG
[Ed. note: an alternative view... (note: Mr. Brick is a well known
Leicaphile and engineer designing autofocus systems etc.]
From Leica Mailing List:
Date: Thu, 27 Jul 2000
From: Jim Brick jimbrick@photoaccess.com
Subject: [Leica] Re: The quintessence of Leica photography? - Long
response -
Erwin,
I agree with you completely. My daughter and I will be standing with you,
with our film cameras, high above Hong Kong (or anywhere else) while the
world has forgotten the craft of photography.
The "craft" of photography cannot be practiced or duplicated with a
scanner, Photoshop, and an inkjet.
Period!
As I've said before, I have a transparency of a field of calla lilies
surrounding an old decaying wooden fence. This image can only be printed
on Cibachrome (Ilfochrome). I've tried to make a LightJet print but to no
avail. The black under and surrounding the plants is like an abyss. And on
supergloss, it looks wet and deep. The green leaves have an electric glow
along the edges. The white lily faces have delicate detail within them.
This is a dynamic range that is stunning in a 30x40 print, the deep deep
abyss black and the delicate white lily faces, plus the glow of the
leaves, but simply "cannot" be reproduced digitally even though the
LightJet printer prints on photographic paper.
And my local lab (Calypso Imaging) just quit printing Cibachromes. It's
either RA-4 or LightJet now. I can print Ciba's up to 20x24 in my own
darkroom but I currently have a order for some 30x40 Ciba's (the calla
lily image) and I now have to drive to San Francisco to get them printed.
Looking at test strips is very inconvenient. This time I'm going to have a
dozen printed so I don't have to go back as often.
So even in the pro labs, the work process is shifting toward digital. Part
of the reason is that the pieces of equipment to produce pro level digital
work are outlandishly expensive. $250,000 - $500,000 for a LightJet
printer. $100,000 and up for a good drum scanner. These pieces of
equipment have to be busy nearly 100% of the time in order for the lab to
stay afloat. Especially since these "state of the art" pieces of equipment
are only state of the art for a couple of years. Then it's buy it all over
again. So the work effort is shifted toward getting digital customers.
Lots of digital customers.
My daughter, who is 20, is majoring in photography and music in college.
They are teaching the "craft" of photography from the ground up. Real
silver photography. Last semester was the zone system and they had to use
D76 1:1 and could not use Delta films as they do not respond linearly to
the zone system expansion and compression techniques. Ilford FP films,
Plus-X, Tri-X, APX 25, APX 100, etc. Real old fashioned silver film. My
daughter uses APX 100 and her prints, on Ilford FB WT, are gorgeous.
Brilliant sparklie highlights and the tones slide from sparklie white into
a deep black that suck you right in. This is only available on wet
processed photo paper from negatives that have been exposed properly and
developed properly based upon the dynamic range of the subject and how
you, the photographer, visualize the resulting print.
This process of visualizing a result before the image is captured is a
silver halide process. Learning the craft of photography teaches you to
view your surroundings in terms of a final print. Your technique takes
into account all of the variables within the scene and, using that
magnificent gray matter computer, exposes correctly, in terms of how the
film will be developed, and in terms of what kind of paper it will be
printed on. This is not a simple process and can only be learned with
practice and many mistakes.
This is not usually the case with a digital camera or even a film image
that is going to be scanned and inkjet printed. The process of
visualization of the final print most likely takes place in Photoshop.
While learning the zone system, my daughter was out in the forest
photographing some tree scenes for her class portfolio. It was dark under
the trees, very bright in the open space behind the trees. Dirt, rocks,
dry grass, a trail running through the scene, etc. Normal forest stuff.
She set up her Hasselblad for a particular scene, used a spot meter to
meter the various important subjects, visualized how she wanted to final
print to look and figured which subject zone to place where on the scale
and how to process the film. N+x, N, N-x. She chose the back for that
particular development time and photographed the scene. THEN... she took a
back that was not a zone specific back, put it on the camera and used the
built-in camera meter to simply photograph the scene. Just like anyone
normally would. She did this with all of her portfolio photographs as sort
of a reality check.
Back home she processed the film (APX 100 in D76 1:1) from the various
backs at the appropriate times that she had worked out when she calibrated
her procedures to the zone system. She also developed the non zone roll at
the normal APX 100 - D76 1:1 time. All of the negatives looked great. Even
the non zone roll. They were just good healthy looking negatives. Then she
started printing.
She first printed her favorite scene from the non zone roll. The print
(11x14) looked good. A little dodging and burning here and there, but a
reasonable print. THEN... she printed the same negative from the zone
roll. She nearly fainted. She came out of the darkroom yelling "DAD...
look at this!" A straight print that was so much better than the non zone
print, it was stunning!!! The tones slid from bright sparklie white into a
deep seductive black. The difference between the two prints was simply
amazing.
This folks, IS the "craft" of photography. It is not simple. It is not
"point and shoot." It requires visualization and thought. It requires a
thorough knowledge of the processes involved. It requires work, which is
where many people give up.
The digital process has solved this for those folks. Simply point and
shoot. Scan if it's not already digital. Fix-up and manipulate in
Photoshop. Print a pleasing inkjet. No photography craft involved. Just
move the pixels to where they look good and be done with it.
All of you real photographers out there, those versed in the "craft" of
photography, should make it a life long commitment to pass on your
knowledge, get a young person involved in silver based photography and wet
darkroom work. My daughter, who is a computer whiz, recognizes with little
effort that there is no comparison between a silver darkroom print and an
inkjet print. The darkroom print wins hands down.
Her "minimum" print size is 11x14. When spotting these prints, with your
nose an inch from the print, you can see the crisp image edge sharpness
and fine detail that is non existent on ink jet prints because of dot
bleed and scanner ICE algorithms. Also, not many folks print ink jet
prints larger than 11x14. And the rubber meets the road when you get to
20x24, when the sharpness, fine detail, and dynamic tonal range, just
leaps off of the print. The big inkjet printers use a larger ink dot
therefore close-up inspection of a large inkjet is not advisable.
The craft of photography, can be done at home, with minimal darkroom
equipment expenditure. And the equipment can easily be useful and
producing exemplary work over a lifetime.
I'm happy that my daughter has chosen to learn the "craft" of photography.
She just got engaged two months ago. She and her bo are talking about
buying a house. The criteria, she says, is that it have a good music room
- a place to teach piano lessons, and a good place to build a darkroom. So
at least in my family, pixels will not replace silver halide molecules on
neither the source (film) nor the destination (paper.)
Another generation carries it forward.
Jim
PS... this is not a denigration of those folks that have no possibility of
having and using a darkroom, and therefore are forced to go digital. I
feel
for them and would indeed go that route myself, if I were forced to.
Erwin Puts wrote:
>The seemingly relentless march of digital printing does signify two trends.
>First of all a loss of knowedge of true and important photographic
>principles.
Date: Mon, 07 Aug 2000
From: mos topher mostopher@hotmail.com
To: rmonagha@mail.smu.edu
Subject: http://www.smu.edu/~rmonagha/mf/scanners.html
this message is for the following page, im not sure how to post.
Has anyone had any experience with the Spot Fototak6? It is a 4.5x6"
print, slide, and negative scanner at 600x1200dpi. It sold in '97 for
@200$ and now they can be found for less than $50. I just purchased one,
and am waiting to see what it can do. It came with a propietary ISA card
for my PC. It has no power cord, as it uses power from the PC connection.
I have read that it is very slow, but this does not matter much to me- it
can scan four 6x6 slides at a time. I'm expecting the worst, but hoping
for some good results. It comes with a lighted lid so no adapter is
needed. After I give it a go, I'll let you know how it is if you want.
Regards,
Mos Topher
Newsgroups:
rec.photo.equipment.large-format,rec.photo.equipment.medium-format,rec.photo.tec
hnique.misc
Date: Tue, 01 Aug 2000
Subject: HP $299 scanner does medium-large format negs, etc.
From: "John Stewart" spambot@us.gov
HP just sent me a $299 (list price) scanner that comes with a tranny
adapter for up to about 7x7 inch negs, trannies, etc.
I can gang up 4 6x6 negs in one scan, then select each one for scanning.
1200 dpi is good enough for me to see the images, make some home prints
and do scanned "contacts.
Anyone else used this scanner?C7690B? USB and paralle. Mac or PC, Masks
for 35mm and mounted slided included.
Not enough dpi for pre-press I guess, but great for hobbiests needing to
check out negs and make modest prints.
John
From Rollei Mailing List:
Date: Mon, 21 Aug 2000
From: pkkollas@gorge.net
Subject: Re: [Rollei] Re: Medium forfmat comeback??? APS, scanners
i use a UMAX Astra 2200 that provides for 8-1/2 x 11-1/2 reflective
scanning (prints, documents,etc) and up to 4 x 5 transparencies and
negatives. Cost is/was around 150 - 175$. Works great for any normal CRT
display.
paul
Michael Levy wrote:
snip
> *THIS LEADS TO the question of the day:
>
> I want to scan some 127 negs, 120 negs, possibly some color prints and then
> do as per above and I do NOT want to spend the retirement fund on a scanner.
From Panoramic Mailing List:
Date: Tue, 22 Aug 2000
From: "Xiong, Zonghou" Zonghou.Xiong@dem.csiro.au
Subject: RE: Flatbed Scanning
To avoid Newton's rings with some flatbed scanners put the emulsion side
down on
the glass and flip the image after scanning.
Zonghou Xiong
...
From Panoramic Mailing List:
Date: Mon, 21 Aug 2000
From: Paul Weil pdweil1@earthlink.net
Subject: Re: Flatbed Scanning
you wrote:
>I have been using an Epson 1200 flatbed scanner with transparency adapter for
>negatives and slides up to 4x5, and occasionally an image that will not lay flat
>on the glass. Does anyone have any suggestions on ways to flatten the image?
>Earlier I thought someone on the list may have mentioned anti-newton glass. Is
>this effective? Where is it available?
>
>Allen Lefebvre
>Canada
I made negative holders out of a thin sheet of black mounting board. Works
well, and is cheap.
Paul D. Weil
pdweil1@earthlink.net
Date: Mon, 25 Sep 2000
From: "Jon Harris" jon_harris@geocities.com
Newsgroups: rec.photo.film+labs
Subject: Re: 200 dpi limit for prints? Re: Film resolution to digital
equiv
I did my own tests a while back with a print made from a point and shoot
camera hand-held with flash on Kodak Gold Max 400 film, printed on Kodak
Edge paper by Qualex. I found that there was definite details gained by
scanning at 300 instead of 200 dpi, but that the gain from 300 to 600 dpi
was negligible. From this I concluded that the maximum DPI I should scan
at
lies somewhere between 300 and 600dpi. Other tests may prove different
maximum DPI, depending on the quality of the print. It stands to reason
that a better print (better paper, sharper image) may require higher DPI
scanning to get the most out of it.
...
Date: Thu, 12 Oct 2000
From: Ernst Dinkla ernst@dinkla.demon.nl
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Convert DPI to lpmm?
john
URL:mailto:stafford@vax2.winona.msus.edu wrote:
> "Q.G. de Bakker" qnu@worldonline.nl wrote
> > john wrote:
> >
> > > Is there a rule of thumb to convert DPI to LPMM?
> >
> > DPI / 50.8 = LP/MM
>
> Thank you! So, does any digital printer produce 4000dpi (79lpmm) prints?
> If it did, there would be justification for MF scanning. The results would
> match the better (perhaps best) of traditional MF enlargement printing.
There's a consensus that for output on Epson inkjets at 720 dpi you get
a significant quality increase up to 240 ppi input, above that it is
less significant but some claim better quality at 360 ppi with 720/1440
dpi printing. With newer Epson models (minimum 3 picolitre drop, 3 to 4
dropsizes, 2880 dpi) the best output will ask for more ppi.
300 ppi will be good enough for any job and most eyes I think.
24 x 36 mm at 4000 ppi scanned delivers a good 30 x 45 cm picture.
56 x 56 mm at 4000 ppi scanned delivers a good 75 x 75 cm picture.
56 x 86 mm at 4000 ppi scanned delivers a good 75 x 115 cm picture.
4000 ppi up to 6 x 9 will be possible with the new Polaroid scanner,
price of that scanner is near the $ 4000 mark I believe.
The Agfa T2500 + a similar Microtek flatbed scans at 2500 ppi,
4 x 5 and panorama, price near $ 3500 for the Microtek.
The Epson 1600U (don't use the 1600S) 1600 ppi flatbed
scanner is a lot cheaper but the ratio to the 4000 dpi Microtek and
Polaroid 35 mm film scanners means that a 6 x 9 film at 1600 ppi will
give almost the same 30 x 45 cm picture as a 35 mm scanned at 4000 ppi.
Given some MF camera prices 4000 $ for a scanner shouldn't be a problem.
But if you have a lower budget and you want to have digital output it
will be difficult to get a better quality than 35 mm delivers.
Ernst
--
Ernst Dinkla Serigrafie,Zeefdruk
From Contax Mailing List:
Date: Thu, 19 Oct 2000
From: Bob Shell bob@bobshell.com
Subject: Re: [CONTAX] Digital vs Film
You are right. I've seen all sorts of numbers thrown around
but in reality there is no direct comparison possible. Even
if you could compare pixels to film grain you would still be
ignoring latitude, which in film far exceeds imaging sensors.
Bob
> From: muchan muchan@promikra.si
> Organization: ProMikra d.o.o., Ljubljana
> Reply-To: contax@photo.cis.to
> Date: Thu, 19 Oct 2000
> To: contax@photo.cis.to
> Subject: Re: [CONTAX] Digital vs Film
>
> But... the information of "one grain" of film and the
> information of "one pixel" is not equivalent. I don't
> know how it can be compared...
>
> muchan
From Panoramic Mailing List:
Date: Wed, 25 Oct 2000
From: CharlesofSC@aol.com
Subject: Re: XPan image scanner
I use my Microtek 4 with the spacers cut out of the seven strip trans/neg
holder. That will cover a full 360.
I just bought the Photosuite Ver 4 Platinum at Staples (39.95 less $10.00
mail in coupon) and that stitches so smooth I can't believe it.
CharlesofSC
[Ed. note: Mr. Brick is an engineer designing autofocus camera systems and
an expert photographer, photobook author, and expert on Leica,
Zeiss,...]
From Hasselblad Mailing List:
Date: Tue, 17 Oct 2000
From: Jim Brick jim_brick@agilent.com
Subject: Re: Digital quality
This is old news and it is not even close to equaling or overcoming 35mm
or 6x6 resolution. A 100mb file is required to archive a 35mm slide. That
means you can take the file and through computer film output, create a
slide that is, for the most part, equivalent to the original file. It
never is as good, but good enough for most purposes. 6x6 requires nearly
300mb and 4x5 is in the stratosphere.
You should read the details of the way Foveon produces the high pixel
count. Both expensive and been around for some time. The Nyquest effect is
still present in any system that has evenly spaced pixels, especially when
they are between 3 and 5 microns in size plus a couple of microns between
them. This is a physical limitation that film does not have and will
plague digital technology until some other technology is invented to
eliminate the individual phototransistor/capacitor/lens pixel circuit. You
cannot make pixel circuits any smaller because there is not enough space
for a meaningful amount of electrons to be held, then read-out, with which
to represent the light level. In other words, the signal to noise ratio
goes to hell and the resulting image is crappy.
Kodak/Foveon uses three
sensors R, G, & B and a prism arrangement to split the image into three
parts. Not new. And the pixel size is still the same as before although no
longer reduced in count by the bayer pattern.
Scanning digital backs
basically do the same thing but it takes three passes, so the subject
cannot move. This will allow a similar resolution but with a moving
subject. And it still requires an umbilical tied to a computer. A local
storage device can be used but the whole thing is large and different. It
will be very useful in many photographic situations but is certainly not a
panacea and certainly will not give film any competition in most of the
areas where film is still in command. And, as I said, does not approach
film resolution capabilities in films like Kodachrome, Velvia, Provia,
Astia, E100, E200, etc.
Jim
...
Date: Wed, 20 Sep 2000
From: spam-abuse@worldnet.att.net (Tom)
Newsgroups: rec.photo.film+labs
Subject: Re: Film resolution to digital equivalent?
-- jan AT bytesmiths DOT com [remove .gov]) wrote:
>TheCharlie@erols.com wrote:
>...snip...
>If a film is rated 80 line pairs per millimeter, you can accurately
>reproduce an 80 line screen by sampling at or above the Nyquist frequency,
>which is two times the highest frequency sine to be reproduced. That's 160
>lpmm, or 4,064 samples per inch (spi).
>
>For example, if your goal is to accurately sample and reproduce the GRAIN
>of the film, you will, of course, have to sample at a much higher
>frequency than if you merely wish to accurately sample and reproduce the
>INFORMATION that the grain is capable of accurately representing. Note
>that to accurately sample and reproduce a perfectly sharp black-white
>transition will require an infinite number of samples, which is where the
>"you can't do that" folks like TheCharlie are coming from.
>: Jan Steinman -- Jan AT Bytesmiths DOT com
>: Bytesmiths -- digital artistry http://www.bytesmiths.com/Art_Gallery
There may be yet another issue confounding discussions of sampling
frequency vs film grain size that has not yet been mentioned.
I believe that most consumer film scanners do not change their optical
sample area when you (use the software driver to) change the number of
samples per mm. Most of these scanners simply keep their optical
resolution maxed out, and take samples further apart.
With such a scanner, if there is a lot of meaningless high spatial
frequency content to the image being scanned (for example, the edges
of film grains), when the samples are spaced far apart, some samples
will land on the center of a grain whereas other samples will be taken
right on the edges of grain particles.
Thus, the variance of the set of sample values from a particular piece
of film will be larger (potentially much larger) than if the optical
sampling area was adjusted to match the sample spacing. Photographers
will see this as an unexpected amount of "noise" in the sampled
image, and will naturally come to the conclusion that (for their type
of equipment) they have to use a very large sampling frequency to get
a good image. What they may not realize is that this frequency is
much larger than they would otherwise need in a scanner whose optical
sample size is adjusted to match the sample spacing.
Their conclusion will be correct for this type of scanner, but a lower
number of samples per mm might produce a final image of the same
overall quality in a scanner where the sampling area could be varied,
particularly in the common case where camera shake, focus, DOF, etc.
is limiting the resolution of the image presented to the film.
Another very related issue is whether the optical system of the
scanner is actually measuring the local absorption of the film being
scanned (ie, the quantity most directly related to the intensity of
the light that formed the image ), or whether the scanner is measuring
some combination of the local absorption and local scattering (say
from edges of the film grains / dye clouds). The latter case will
occur when the f-numbers of the optics on either side of the film are
high (ie, small lenses, highly collimated light source). This effect
has been seen for decades as evidenced by the old debate between
condensor and diffuser light sources for enlargers.
In the case of scanners, high f-number optics (in the scanner) will
increase the amount of meaningless high spatial freqency signal
recorded by the scanner.
Comments?
Tom
Washington, DC
Date: Wed, 20 Sep 2000
From: "Ray Smith" raysmith@callnetuk.com
Newsgroups: uk.rec.photo.misc
Subject: Re: Digital Equivalent resolution of film
Hi Robert,
Great link!
The figures tie in with the kind of figures that Kodak gave me (see post
further down).
For medium speed film, 54 megapixels at 24 bits per pixel comes out pretty
close to 150Mbytes, and Kodak quoted 150 megapixels for slow speed slide
film, at 6 bytes per pixel (actually 42 bits per pixel).
I haven't been able to digest all the tech details yet, but I've
downloaded the article for leisure reading ;-)
From a computing point of view think what the effective "data transfer
rate" is when you can capture 720 Mbytes of information in say 1/250th of
a second using conventional film! Also means that each page of negatives
that I have on file would require 36 x 150 Mbytes per frame or about 5
Gbytes of computer storage in raw form. Multiply that by 3000 rolls of
film per year for n years that I've been working, and I think I've run out
of zeros......... [g]
Cheers, and thanks again for the link
Ray
> see http://ipas1.afip.org/~oliver/wendy/photo.html which gives table:
From Leica Mailing List:
Date: Mon, 30 Oct 2000
From: "Erwin Puts" imxputs@knoware.nl
Subject: [Leica] Film megabytes and some curious contradictions
I am surprised. A few years ago I gave some figures on this list and
noted that Kodachrome 25 had a byte count of 135 MB which is an amount
digital capure would not reach in eons. At that time I got ridiculed and
some even would me have ostriced from this list as being The Idiot of
Tradition. Even recently my ideas and facts about the level of image
quality that can produced by film and Leica lenses have been questioned to
the extreme. Now I present details from "conventional wisdom" (not my
ideas, but industry figures) and again I am approached as the person who
should do his numbers again.
What is the amount of bytes for film. Some say you should count the
grains, but that is the wrong apporoach, as the individual grains are very
small, often in the region of less than one micrometer in diameter. In
practice, the smallest image point needs a number of grains to become
visible, so this does not count. The best equivalent is the resolution,
that is a square grid superimposed on the film area, as this is the exact
replica of a sensor grid. Now current film (BW or slide or color neg) has
resolution figures around 100 lp/mm and Techpan has 200 to 300 lp/mm.
Follow this: the negative area has 24 x 36mm. Every mm holds 200 different
lines. that is 24 x 200 x 36 x 200 bytes = 4800 x 7200 = 34.560.000 bytes.
We need three colours, that is 34.560.000 x 3 = 103.680.000 thus more than
100 MB for a colour negative or slide.
This figure is close to my original and ridiculed calculation. Now for
some steps. It is very difficult and often considered unnecessary (pace
Mike Johnston) to get on film more than 40 line pairs or 80 lines. And if
we be even more general, 20 lp/mm are the best. most people would dream
about. Assume now 20 lp/mm or 40 lines per mm (the best you can get in
hand held picture taking, generally speaking). Again: 24 x 40 x 36 x 40 =
960 x 1440 = 1.382.400. Three colours would be 3 times this number, which
is 4Mb. And with 40 lp/mm we et 5.529.600 bytes times 3 = 16.588.800 bytes
or 17 Mb, quite close to the 20 MB I quoted as the conventional wisdom.
So any digital capture of 20Mb would be close to the resoluton of 40 lp/mm
that some on this list would consider as the most one would want in 35mm
photography and to go beyond this number would qualify as being a freak
obsessed with lines and test targets and all that. Reread the mails by Mr
Johnston and Mr. Grant and Mr Goodman at all.
So I accept this proposition(as does the industry as a whole) and I start
from these figures (20Mb) and now I am questioned by some others, who say
that my figures are wrong and that I should start with a much higher
number
of Megabytes to do justice to current thinking.
Now I am totally at a loss.
Some would hold that 40 lines is the best you need in typical Leica
photography, which amounts to a 4MB picture, easily reached by todays
digital cameras of 3.3 to 4 Mb.
Look at the facts: I noted earlier that you should aim for the 135MB
possible with Kodachrome or even better hi-res BW film. I am shot dead for
this proposal, as it exemplifies The Idiots Approach To Ridiculous HiRes
Photography.
Then I note that maybe 20Mb would be a sensible compromise and I am again
shot dead because film can handle 100Mb or more and now it exemplifies The
Idiots Approach to Ridiculous Claims by Digital Photography.
I am confused!
Can anyone explain?
Erwin
From Hasselblad Mailing List:
Date: Tue, 14 Nov 2000
From: olenberger olenberger@email.msn.com
Subject: Re: Scanners...
Rich:
The Epson will scan up to 4"x5" transparencies, has a resolution of
1600 x 3200 dpi, 42-bit color, and DMax of 3.2.
The HP will scan up to 5"x5" transparencies, has a resolution of 1200
x 2400 dpi, 42-bit color, and DMax that is unspecified (I e-mailed HP to
get this spec, and they responded that they did not specify DMax, and that
other things such as quality of design were more important).
The UMAX will scan up to 8.5"x10" transparencies, has a resolution of
1200 x 2400, 42-bit color, and DMax of 3.4.
In my research of scanners, the value of DMax, which tells you the
maximum density of film the scanner can render before it gives up and
calls it black, appears to drive the price. The scale is logrithmic, so
there is a lot of difference between 3.0 and 3.4, or so they say. I
believe a value of 4.0 is perfect -- or close to it -- although you would
have to go to a very expensive drum scanner to approach that.
We have all seen the contrast gain that occurs when we make prints
from slides, so that is why DMax is important, especially if you have some
contrasty slides.
I would think color depth (42 bits in all cases) would relate to DMax,
but the specs on the various scanners don't seem to bear that out. I
think the source of light has more to do with it.
All these specs probably don't answer your question "are all three
decent?" I personally don't have a feel for the difference between a DMax
of 3.0 versus 3.4. To really assess that, you would have to scan the same
contrasty slide on all three scanners and look at the difference in the
output. The service bureau I have been using has samples output from a
Nikon 35mm film scanner, a flatbed scanner, and a drum scanner. And yes,
there is a difference. And they charge accordingly.
Hope this helps.
-Fritz
...
From Nikon Mailing List:
Date: Wed, 22 Nov 2000
From: Adam Pierzchala adam@pierzchala.newnet.co.uk
Subject: [NIKON] F of human eye
Tyler wrote: "I seem to recall while reading something about the Agfa
high speed film innovation that it takes (I really do not remember) 3-5
photons of light to hit the SAME grain/crystal before the film ever
records anything"
Well, I think your memory is spot on. When I used to work at Kodak, the
wisdom of the day was that 4 photons were needed to kick the grain into
action.
The flat T-grains which are like platelets ensure that a large surface
area is presented to light, so that there is a better chance of those
photons actually hitting their target. This is better than having a
narrow edge of the grain towards the lens.
I must admit though, that these grains are vastly larger than any
photon, in whatever uncertain state it happens to be, so why we need
acres of surface area is beyond me... :)
Regards, Adam
Panoramic Mailing List;
Date: Tue, 07 Nov 2000
From: Mike Sinclair sinclair@microsoft.com
Subject: RE: Digital Cirkut Cam?
Scott,
The best "cheap" scanner I have found lately that does a good job is the
Epson 636U (~$90 street). It has a 5100 element trilinear CCD, internal 12
bit digitizing, 600dpi intrinsic resolution and USB interface. This will
yield a maximum raw file size of ~105Mbytes. Also, the sensor is 42mm
which is ideal for digitizing the largest image circle from most 35mm
lenses.
There are three (at least) problems with this approach (and applies to
most of the scanners on the market now). One is most of the inexpensive
scanners have little or no built in gamma compression. Typical contrasts
on things that flatbed scanners scan is usually less than 100:1 but if you
point the lens/sensor at the world, you'll get many orders of magnitude
beyond this. Another problem is calibration. Used to be most scanners
would scan a calibration sheet, store the result in the computer then use
it for every subsequent scan. New scanners calibrate by scanning a 8.5"
white strip before every scan. This requires you have an equivalent
calibration setup in your camera. Another problem is that most flatbed
scanners have fixed maximum scan lengths, usually around 11.5". Even if I
use a Zenit 16mm fisheye (www.russia2all.com -$139), I need a longer scan
, ~17", to give me ~square pixels, though having to change the scaling is
not that bad. This setup gives me ~180 x 360 degree scans.
Good luck and happy scanning.
-Mike-
From: sperry@mainewest.com [mailto:sperry@mainewest.com]
Sent: Tuesday, November 07, 2000
To: panorama-l@sci.monash.edu.au
Subject: Digital Cirkut Cam?
While thinking of a winter project to do, I'm considering the feasibilty
of putting parts from a one pass CCD flatbed scanner in a homemade back
that will fit on a Cirkut camera outfit. Don't ask me why, but any
suggestions on how this could be done would be appreciated? A year or so
ago I remember a fellow on this list was building a digital panoramic
camera with, if memory serves me correctly, a Marstek hand held scanner.
Are you still on this list? If not does anyone else have his email
address?
-Scott
-Scott Perry
-sperry@mainewest.com
From MEdium Format Mailing LIst;
Date: Wed, 03 Jan 2001
From: Schatzie Walton jdwalton@home.com
To: medium-format@egroups.com
Subject: [medium-format] Re: Scanner advice needed
The EPSON 1640SU + Photo will scan 35mm, 66 and 45's - it's the
highest resolution unit at its price point ($399 retail with the
transparency adapter). Bundled with the scanner is Photoshop LE,
Photo DeLuxe and EPSON's proprietary software. Uses USB port and it's
pretty quick.
I doubt that it will do the job of the kilobuck models, but horses for
courses. I just got one to replace an HP with broken transparency
adapter. HP is not very good on aftermarket support - they wanted
$1200 to replace the transparency adapter!
[Ed. note: check the new nikon medium format scanner ($3k?) as well as the
Polaroid printscan below:]
From Hasselblad Mailing List:
Date: Wed, 3 Jan 2001
From: "Q.G. de Bakker" qnu@worldonline.nl
To: mikec@america.net, hasselblad@kelvin.net
Subject: Re: More interest: Polaroid Sprintscan 120
M. Callahan wrote:
> Count me in as another person extremely interested in hearing about the
> Polaroid Sprintscan 120. Not sure that it is even available yet, but is
> promised in the first quarter of 2001. Has anyone heard anticipated
> pricing yet?
>
>From Polaroid's site:
"The SprintScan 120 will be available in the United States in January 2001
through authorized Polaroid dealers and catalogs. The suggested U.S. list
price of the new Polaroid scanner is $3,995. International prices may
vary, and can be determined by contacting the regional Polaroid office. To
obtain more information on the SprintScan 120 or other Polaroid scanners,
customers may call 1-800-816-2611, extension D267, or access the Polaroid
business-to-business Web site at http://www.polaroidwork.com/."
From Rollei Mailing List;
Date: Tue, 16 Jan 2001
From: Bob Shell bob@bobshell.com
Subject: Re: [Rollei] Scanner question for MF
Wait a few months. Nikon's new 8000ED should be on the market then.
Under US$ 3,000. High res, Digital ICE, takes 35mm as well, etc.
Sounds like the "dream machine" to me.
Bob
rec.photo.equipment.medium-format
From: "alex wetmore" alex@phred.org
[1] Re: camera kit recommendations for MF beginner?
Date: Thu Feb 08 2001
....
For $1k you can buy a flatbed scanner for MF negatives and a film
scanner for 35mm film. I have an Epson 1640SU that I've used for
scanning MF film (it works great!) and a Nikon Coolscan LS-1000 that I
use for 35mm film and slides.
The 1640su gives me 3600x4200 pixels for 6x7film or would give me
3600x2700 pixels for 6x4.5 film (I don't have any 645 images, but plan
on buying a 6x4.5 camera in the next few months). The LS-1000 gives me
about 3800x2400 pixels for 35mm film. I find that MF is a lot easier to
work with when scanning because the dust spots aren't as noticable or
large and film grain isn't a concern. Any of these resolutions is
enough for decent 8x10 prints on a quality photo printer.
alex
From Leica Mailing List:
Date: Thu, 1 Mar 2001
From: "Stanislaw B.A. Stawowy" watteau@krakow.neurosoft.net
Subject: Re: [Leica] Scanning tutorials
> Also, is there a consensus on good alternatives to Adobe Photo$$$hop?
> For what I want to do, Photoshop seems like very expensive overkill.
> How much does one lose using Photoshop's "Lite Edition." If I went with
> another program, would I be depriving myself of the vast knowledge pool
> regarding Photoshop?
Simplest and most powerful one is Irfan View: http://www.irfanview.com
(nearly_) free (you have to send a postcard) program, aimed at
least at muy needs perfectly. I still use Fireworks and PS also,
especially at work, but Irfan is simply faster and easier.
- -----
St.
VF750 - 'Corvus'
(Stanislaw B.A. Stawowy)
http://www.geocities.com/Stanislaw_Stawowy
From Leica Mailing List:
Date: Thu, 1 Mar 2001
From: "Dan Post" dpost@triad.rr.com
Subject: Re: [Leica] Scanning tutorials
Peter-
check this site-
http://come.to/digitaldarkroom
They have all kinds of info and links for scanning and printing digital
images, and a wealth of data on printing with Epson printers which seem to
be a favourite here on the LUG.
Good luck!
Dan
From: flexaret2@aol.com (FLEXARET2)
Newsgroups:
rec.photo.equipment.medium-format
Date: 29 Mar 2001
Subject: Low
Cost Film Scanner
As you all know a good scanner for Medium Format Negatives and
Transparencies costs at least $2000 - ie. the Minolta Scan Multi.
But, for those on a low budget, what can you get for $79-$99.
- The Umax Astra 2200 - now being closed out for $79 to $99. Those without
a Scanner for Medium Format film - grab it - if only for experiments.
I just got one and I am amazed. This flatbed has its own built-on
Transparency Adapter (Illuminated Cover) which will cover up to 4" x 5".
The scanner will work with MAC or PC - USB or SCSI.
I just got mine in today and am amazed with the good quality I got with a
2 1/4" x 2 1/4" Color Transparency and a similar sized Black and White
negative. I had no luck with a color negative, but there must be some
trick to getting rid of the orange mask - or I made need other software
than Vistascan and Photoshop 5.0 LE (Limited Edition) which comes with the
scanner. Optically the quality is quite good and I intend to do extensive
tests with it.
For my purposes I just need scans for small video box cover design, but I
know that good 8x10 prints are possible as I printed one out from the B/W
negative.
From Leica (Topica) Mailing List:
Date: Wed, 28 Mar 2001
From: Jim Brick jim@brick.org
Subject: Re: Kodak Makes Improvements to B&W; Films
The problem with consumer film scanners is that the resolution (dpi) is in
the neighborhood of the film grain dpi, depending upon the film of course.
Films like Tech Pan, Velvia, Kodachrome 25, APX-25, etc, have a grain
structure finer than the scan frequency.
The grain dispersion frequency is higher than the scan frequency. A
classic case of the Nyquist limit. You cannot capture high frequency data
using a low frequency capture mechanism. Digital lab instruments run into
this all of the time.
There is no easy way out other than buying a higher dpi scanner. Basically
a drum scanner or equivalent. Software interpolation routines can
interpolate out the fringing, aliasing, and other artifacts, but you still
are not capturing what you really have in the transparency or negative.
Only an interpolation of it. Look up "interpolation" in the dictionary.
The other way around though, scanning Tri-X developed in Rodinal, etc.,
will fairly accurately record the film detail, grain and all. This is
because the image make-up is of a lower frequency than the scanner
frequency. You get the grain accurately represented. Which should be what
you want as this IS film that you are scanning. If you interpolate out and
smooth out the grain, you again are creating something that you didn't
have originally. Not wrong, just different.
This is why I still don't own a scanner. When I scan something, I have it
drum scanned producing a 300mb file. This is good enough to produce
LightJet prints up to the limit of the LightJet printer. 50"x96" And the
grain of the film is accurately depicted. I like this.
Someday I will get a used Ima??? (Imacon, Imatek, ??????, whatever the
name of it is) and be able to scan my film at up to at least 12,000 dpi.
Since Iprint most of my work on Cibachrome up to 20x24, my need for
scanning is not very great. There is a lab 40 miles away that can print
Ciba's up to 40x60 so I use them for larger than 20x24 Ciba's. Or have
Calypso do a scan and have a LightJet made. Either way, my personal need
for a scanner has not arrived yet.
The average size of a digital camera pixel is 5 microns. It takes four
sensor pixels to record one color pixel. Leica lenses record finer detail
than what can be recorder with five micron pixels times four (100 sq.
microns.) You can have many pieces of fine detail falling onto this 100
square micron pixel area, which cannot in any way be separated out. Again,
capture frequency less than the data frequency. And if the detail density
happens to equal the pixel spacing density, you can miss a whole lot of
data. This is why resolution limiting filters must be used with consumer
CCD or CMOS sensors. These sensors are simply incapable of capturing the
detail that the lens can provide. And as a result, all kinds of strange
artifacts occur. Many commercial and scientific sensors are on a mechanism
that will shift the sensor slightly and take four images, then software
reconstructs the image, putting back together the image at the intended
resolution. Or like the Leica S1, a very high resolution single array CCD
was scanned over the image, in minuscule steps, producing a similar
result.
Being able to capture (almost) Leica lens resolution.
From Rollei Mailing List;
Date: Mon, 11 Jun 2001
From: Michael Levy yvel@adelphia.net
Subject: Re: [Rollei] Re: Rollei OT: Scanners: Polaroid Sprintscan 120 vs
Nikon 8000
You don't say what you want to do with the sans...if you don't need huge
gallery-quality enlargements a good place to start is the Epson 1640 SU at
around 400 bucks. It can scan up to 4x5 negs and scan prints or documents
up to legal pad size.
It is not the best -- no flatbed is, when it comes to scanning negs, but
it is pretty darn good and will teach you enough and be challengi ng
enough so that when dual purpose scanners come down in price (they are
about $2000 and up now) you will know exactly what yoyu need and what you
want yor scanner to do.
This one works okay scanning negs for prints up to 8x10 inches with a good
inkject printer if you use high resolution scans
Mike levy
Date: Tue, 19 Jun 2001
From: "Joseph Schutz" jfreischuetz@home.com
Newsgroups:
comp.periphs.scanners,rec.photo.equipment.large-format,rec.photo.equipment.mediu
m-format
Subject: Sprintscan 120 for 6x17 and 6x12 Scans
Hello,
If you are interested in scanning 6x12 and 6x17 scans with the Polaroid
Sprintscan 120 read the following response I got from Polaroid
Corporation.
The have indicated that they are interested in supporting these other
formats, but would like to hear from you requesting this support.
Please read the reply to my question below:
Subject
>Sprintscan 120 and 6x17 negatives
>Suggested Answer
>Short answer: yes we are. We hope to add a user definable length
>mode to the software within 6-7 months.
>What that means is that you will be able to scan 6xN images
>where N is limited to the length of the carrier.
>You understand though, you will be consuming a _LOT_ of
>hard drive space when you do that. You will, in no time at all
>exceed the storage capacity of a standard CD with one image.
>Your dealer was right too. it's an issue with the software rather
>than the hardware.We encourage your telling your NG
>about this. It also will speed up matters if people
>would write us telling us that they want this feature.
>Question
>I would love to buy your Sprintscan 120 product, since I have a Hasselblad
> and shot a fair amount of 6x6. My problem is that I also have a Linhof 617 IIIS,
>which makes negatives 6x17, and I need to scan these as well. I finally got a
>look at your scanner yesterday at Citizen's Photo in Portland, Oregon. The film
> carrier easily would accomodate the 6x17 negative size, and the salesman told me
>he thought it would limited by the software or firmware. I also use my 4x5 camera
>with a 6x12 back, and would like to scan those negatives as well.
>I would like to know if you are planning to support formats larger than
>6x9 any time soon? It looks that your hardware could do it.
>I have been following this topic in the usenet news groups for quite a
>while now, and I tend to think that there would be quite a lot of interest
>in 6x12 from the 4x5 crowd.
From: rafeb@channel1.com (Rafe B.)
Date: Sun, 03 Jun 2001
Newsgroups:
rec.photo.equipment.medium-format,rec.photo.digital,rec.photo.misc,comp.periphs.
scanners,rec.photo.equipment.misc
Subject: Re: Scanning 6x7 film? (DANIEL ROBERTSON RETURNS) :-)
john@stafford.net (John Stafford) wrote:
>daniel_a_robertson@my-deja.com (Daniel A Robertson) wrote
>
>(Daniel was responding to my post snipped here for brevity.)
>
>> At 1200ppi (the limiting scan here) on the 6cm dimension of a 6x7 (6cm
>> x 7cm)neg/transparency, you would get a linear 2625 pixels (remember
>> the "6" is only about 55.6mm). So if output on an A3 color printer,
>> with the short edge corresponding to the same dimension, you would
>> have "only" 225 dpi. And modern printers *CAN* output a good deal more
>> than that.
>
>I want to believe, and I do respect your mathematics, however let me
>take this discussion to a relevant challenge. Your response could
>serve to settle this matter for The Rest of Us. Here's what I want. I
>want to put side-by-side my (nominal) 6cm sq negative put to 60cm sq.
>prints: one scanned at 1200ppi then printed on a consumer-grade 600dpi
>printer and another print from the same negative done on a
>conventional enlarger. Given the negative has a reasonable *80lpi
>resolution and fine-enough grain to be almost indiscernable at 60cm,
>and the enlarger lens has a 80lpi resolution. (Outcome is about
>70lpi). Which is going to yield greater fidelity (more accutance
>sharpness, etc.)? And will a 4000ppi scanner be over three times
>better _on the print_? What printer will show the better results?
>
>(I've put $3,500US aside for a good MF/5x4 scanner and am hanging
>loose until I feel better about this whole schtick. I'd hate to be
>disappointed if I cannot _see_ a _profound_ difference in my prints
>with this new scanner.)
>
>(* I chose realistic lpi figures for the taking lens, film capacity
>and enlarging lens in order to obviate arguments involving these
>metrics.)
For Epson photo printers, the "effective" and optimal lpi is anywhere
from 240 to 480. Most folks would say 240, but a few diehards insist
they can see improvements up to 480 (I have serious doubts, myself,
about the high end of that range.)
My own scans are anywhere from 240 to 360 lpi at the print, and there
is little discernable difference, as far as I can tell.
1200 dpi is, IMHO, *much* too low a figure for scanning film. 2700
dpi gets you 80% of the way there, I estimate. A *true* 4000 dpi
would get probably 95% of the data that film could store, under
optimal conditions (low ISO, tripod, excellent optics, etc.)
As to the difference between 2700 dpi and 4000 dpi, that is more a
matter of specsmanship on the part of the scanner manufacturers.
It's one thing to have a CCD imager with a given number of pixels,
distributed over some width of input media. It's quite another thing
to have an internal optical system, and sufficiently low noise, in
order to deliver the rated resolution.
For example... I've compared scans from a SprintScan Plus (2700)
versus SprintScan 4000. I surely didn't see a huge difference in
sharpness between these two. I have also compared the SprintScan
Plus scans (2700 dpi) against those from a Leafscan (5080 dpi) and
in that case, there is a very discernable difference, in favor of the
Leafscan.
Speaking from personal experience, making 8x10" and 11x14" prints
from 35 mm and 645 negatives -- the results from my "digital darkroom"
are far better (both in sharpness and color rendition) than anything I
could achieve in a "wet" darkroom.
rafe b.
Date: Sat, 21 Apr 2001
From: sinai_gary clubs-mail@yahoo-inc.com
To: rmonagha@post.smu.edu
Subject: Re: Prime 1800-U Film Scanner [Yahoo! Clubs: Classic 35mm
Compacts]
I have had a Prime 1800-U for about a year now. I too really like it. When
I bought it, it was the only inexpensive 35mm film & slide scanner
available. The next best was about $600. It is very good for black and
white and that is what I use it for mostly. I find that I can never get
the colours correct on scanned colour negs. The slides are very good too
but colour is also a problem. With b&w; I am very happy. It makes my
entire b&w; process an at home job. If you would like me to send you a
couple examples let me know.
Gary
Date: Mon, 09 Jul 2001
From: Bill Tuthill ca_creekin@yahoo.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Resample or higher dpi when downsizing?
All very good advice, but nobody has mentioned printer resolution --
you get best results if you print at a PPI value that divides evenly
into printer hardware resolution. On a newer Epson with < 6 pL drops
you get best results at 360 ppi, second best results at 240 ppi, and
a smaller image but similar sharpness at 300 ppi. For proof, see the
link to Epson 750 images at the bottom of this photo.net thread:
http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=0012tY
I'm not sure about other printers, not having owned anything but Epson.
...
From: bhilton665@aol.com (BHilton665)
Newsgroups: rec.photo.equipment.35mm
Date: 09 Jul 2001
Subject: Re: Resample or higher dpi when downsizing?
>From: Bill Tuthill ca_creekin@yahoo.com
>All very good advice, but nobody has mentioned printer resolution --
>you get best results if you print at a PPI value that divides evenly
>into printer hardware resolution.
No you won't. I've printed at 240, 300, 360 or 600 dpi (240 and 360 are
even multiples of my printer rez) with good results but can't see any
difference if I get a slightly odd crop and print at 287.33 or 308.79 or
any other fraction.
Date: Mon, 09 Jul 2001
From: "Tony Spadaro" tspadaro@ncmaps.rr.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Resample or higher dpi when downsizing?
Yep. Currently there are no printers that need more than 300 dpi and most
need less. It does not matter what the output dpi is. Some will argue for
360, and they have a case, but certainly never more than that.
However this may not always be the case. I speak for summer 2001 - not
winter 2002, and certainly not 2003.
--
http://home.nc.rr.com/tspadaro/
From Rollei Mailing List;
Date: Mon, 23 Jul 2001
From: Bob Shell bob@bobshell.com
Subject: Re: [Rollei] Somewhat OT - Anyone know a cheap place to get 6x6 p ut on ProCD?
> From: "Fox, Robert" RFox@aarp.org
> Date: Mon, 23 Jul 2001
> Subject: RE: [Rollei] Somewhat OT - Anyone know a cheap place to get 6x6 put
> on ProCD?
>
> Any good quality flatbed will make good scans, and cleaning up the scan with
> Adobe Photoshop is simple. Flatbed scanning does take some time to do, but
> you save money in that you only need to scan the best prints, not whole
> rolls. I find that I scan maybe two prints per roll, sometimes none, and it
> takes about 10 minutes to scan and cleanup the photo (even at 1,500 x 1,500
> for a 5x5 print from a TLR).
In terms of after scanning processing, the best thing I have found is
iCorrect Professional. One click gets me close to perfect every time. It
is particularly good on skin tones.
Bob
From Rollei Mailing List;
Date: Mon, 23 Jul 2001
From: Austin Franklin darkroom@ix.netcom.com
Subject: RE: [Rollei] MF scanner resources...was...something else ;-)
> what do you use for scanning your medium negs? I know you may be getting a
> new scanner every month from manufacturers looking for reviews, but is
> there one piece of equipment that us, mere mortals, can afford for our
> hobby?
I would suggest checking out this URL:
http://www.halftone.co.uk/
and also joining the filmscanner mailing list, and asking there. There
are also some on-line reviews, but I don't know if there are any of MF
scanners.
Personally, I don't know that there are any new dedicated film scanners
(not flatbed) that are actually affordable, per se. I'd recommend
checking out a used one on eBay, such as the Minolta...lower resolution
than the new crop, but probably good enough for most people's uses.
I have heard that some people do like the results of the, I believe,
1640SU...which is available for under $200...
You also might want to say what you are planning on scanning for...such as
size of output (web, 8x10, 13x19 etc.) and what your audience is (fine
art, just printing snap shots... etc.).
From Leica Mailing List;
Date: Thu, 26 Jul 2001
From: "Peter A. Klein" pklein@2alpha.net
Subject: RE: [Leica] RE: Re: Digital Leica and reality
In the never ending saga of Kingdom of Shannon, yea, verily, from the
Book of Nyquist, Chapter IV, verse 65,535:
>> difference between the 2700, 4000 and 6000 dpi scans
I was just looking at an excellent site which shows some more of this
stuff.
http://www.cix.co.uk/~tsphoto/tech/filmscan/menu.htm
Here are some more samples, of a little corner of a standard Kodak E6 test
slide, which just happens to have a real-world subject on it. Try these
for size. The site author recommends saving the files and looking at them
in a image editing program so the browser doesn't munge things by resizing
on screen.
The test slide, for reference. Note the size of the face on the slide,
which includes all the gray areas:
http://www.cix.co.uk/~tsphoto/tech/filmscan/howtek/howtkq60.jpg
Now, look at just the face.
Drum scan, 4000 ppi, sampled down to 25% for reasonable download:
http://www.cix.co.uk/~tsphoto/tech/filmscan/howtek/htekface.jpg
Polariod Sprintscan 4000, 4000 ppi:
http://www.cix.co.uk/~tsphoto/tech/filmscan/pol4000/4000fac2.jpg
Nikon LS-2000, 2700 ppi:
http://www.cix.co.uk/~tsphoto/tech/filmscan/ls2000/ls2000fa.jpg
HP PhotoSmart S20, 2400 ppi:
http://www.cix.co.uk/~tsphoto/tech/filmscan/hp_s20/s20face.jpg
To my eyes, the drum scan is a quantum leap better than the Polariod
4000 ppi, which is slightly better than the Nikon LS-2000 at 2700 ppi,
which is a bit better then the HP S20 at 2400 ppi.
- --Peter
Date: Wed, 30 May 2001
From: Carey L. Jones careyj@pipeline.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Sadly, I'm at a loss. Please advise...
rafeb@channel1.com says...
> On Tue, 29 May 2001 Carey L. Jones
> careyj@pipeline.com wrote:
>
> >rafeb@channel1.com
> >says...
> >> On Mon, 28 May 2001, Carey L. Jones
> >> >
> >> >The question is whether speed or image quality is more important to you.
> >> >Answer that question, and your choice becomes clear.
> >>
> >>
> >> I wish it were that simple, Carey. The question is: do I go even
> >> deeper into the realm of film -- by investing $3000 in a Nikon 8000ED,
> >>
> >The Coolscan 4000 (35mm) is only $1700, and gives better resolution and
> >dynamic range than any digicam likely to come out in the next three
> >years. Your 35mm lenses will work the way God and the lens designer
> >intended them to work (your wide-angles won't be turned into instant
> >telephotos).
>
> I already have a decent film scanner (Polaroid SprintScan Plus)
> for 35 mm. But it won't handle 120 film. As you may know, that
> sort of resolution (2700, 4000 dpi) is not readily available in
> scanners that can also handle 120 film. If I had a spare $10K
> to throw around, I'd go for an Imacon Flextight. I'm also looking
> closely at used LeafScan 45s if/when the price is right.
For the cost of an SLR digicam witch will match your Polaroid scanner's
quality, you could buy a Coolscan 8000, which will handle anything up to
6x9 film at 4000 dpi and 4.2 dynamic range. I hate to think what you
would pay for a digicam which will match that, even if anybody made one,
which they don't.
> As of about a year ago, I've tried moving into medium-format,
> with a Pentax 645. The scanner I use (Epson 1640 SU) is barely
> up to the job. But it's almost a moot issue, because many of
> my films come back with dust, scratches, fingerprints, etc., that
> render the best images unusable.
Nikon scanners (including the 8000) come with "Digital ICE", a
hardware/software technology which automatically cleans most dust,
scratches, etc. It's able to recognize surface defects, figure out what
*should* be under the defect, and "clone" it in from the surrounding
image. Does a helluva job, too. I rarely have to do much clean-up with
my Coolscan 2000. The little bit it doesn't get, I can fix with the
Photoshop clone tool in a minute or two.
> >> another few hundred in a Jobo processing kit, chemicals, film dryer,
> >> new lenses for my 645, etc -- or just give up on film.
> >>
> >
> >You only need the Jobo for processing color. A dishpan full of water
> >will keep B&W chemicals close enough to proper temperature for good
> >results.
>
> Well, I am in fact talking about color. I've done plenty of BW film
> processing and darkroom work in my time. I don't miss it much.
>
> >> For that kind of money, in another few months or a year at the
> >> outside, I could be looking at a 6 Mpixel digital camera from Canon,
> >> Nikon, Contax, Pentax, etc.
> >Which will barely do a decent 8x10 print, if you don't do any cropping.
> >A $1700 Nikon Coolscan 4000 will give you around 20 megapixels to play
> >with from 35mm film. It *and* a Jobo processor would run you less than
> >your hypothetical (nobody's actually selling one yet) digital SLR body,
> >and if the digicam doesn't match your lenses, you're out another couple
> >of grand for a new lens kit.
>
> The digicam will of course match my lenses. The Fuji S1
> would do that now, or even the Nikon D1. Yes, I am aware
> of the focal-length-multiplication issue. That's one of the
> reasons I've been waiting for digi-cams with large CCD or
> CMOS sensors that approach the size of a 35 mm frame.
>
> My first film scanner (Microtek 35t+) deliverd 6-MPixel images,
> and from that I made some excellent, prize winning 8x10" prints.
> I get about 50% more pixels with the SprintScan (360 dpi at
> the print, vs. 240) but the difference in print quality is fairly
> subtle.
>
> >> I've developed hundreds of rolls of Tri-X in my day, and even a few
> >> rolls of Ektachrome. But that was a long time ago, and I frankly have
> >> no real desire to revisit that era. My wife doesn't like the idea of
> >> chemicals in the house. I'd *much* rather pay someone to process my
> >> film(s) but I've yet to find anyone that will give my C-41 film the
> >> respect that I feel it deserves.
> >
> >Shoot E-6. Most pros do. It's cheaper to process, and most labs are
> >more consistent with E-6 than with C-41. I've had *much* better luck
> >with E-6 than with C-41 at my local labs. I also find it easier to judge
> >a shot from a slide on a light box than either a contact sheet or a
> >machine print. And if you ever decide to process your own, E-6 has a lot
> >more latitude in processing temperature.
>
> Print film has substantially wider exposure latitude and is in fact
> easier to scan. I have no interest in projecting slides. My sole
> interest is in getting my images into digital form, with the best
> quality possible. If I could afford a scanning back for my 645, I'd
> buy it, but those start at about $10-$15K.
Print film is *not* easier to scan. If you use ICC color profiling
software, such as Monaco EZ-Color, you can get transparency IT8 targets
and build a profile specifically for your scanner. Once you create a
profile, you can scan slides and expect the color to match almost exactly
(assuming your monitor is properly calibrated). Unfortunately, they
don't make IT8 negative targets. They can't, because every different
print film emulsion has a little different color balance. You'd need a
set of targets for every brand of film you use. When I scan Fujicolor
with my Nikon, the images come out okay, but when I scan Kodak Gold, it
can come out with a major green cast.
I don't own a slide projector. I doubt I ever will. I have my lab page
my 645 slide film in strips of 4, and just plop the page down on my light
box and check them with a loupe. What I see is what I get, and I don't
have to keep up with contact sheets. If I see one I want to print, I
pull the strip, put it in the film holder for my Epson 1600, shoot it
with a little canned air, put it in the scanner, scan it into Photoshop,
tweak the levels/curves, clean up any dust spots with the clone tool, and
sharpen the image. Color matching/balance isn't a problem, because I
have ICC profiles for everything in my system.
As to the relative quality of slide vs print film, 90% of the color
images you see in magazines and calendars were shot on slide film. Guys
like Galen Rowell, Franz Lanting, and Jack Dykinga are all shooting slide
film. Do you think National Geographic sends Rowell halfway around the
world to shoot lousy film? Editors *like* being able to look at the
original film instead of a print. They want to reproduce from the
original, so they want to see it. The fact the image is viewed by
transmitted, not reflected light helps, too. Slide portfolios are easier
to ship, and less likely to come back with fingerprints all over them.
The one category where print films deliver significantly superior quality
is fast ( >= 400 ) films. The new Provia 400F may have narrowed the gap,
but I haven't tested it, so I can't say for shure.
> I do agree that "pro" processing labs are biased toward E-6 process,
> and that's certainly a big problem with my choice of C41 process.
They are biased toward E-6 because their pro customers shoot it.
> >> As it turns out, I discovered just yesterday that a local lab that
> >> I've been using -- one of the few that was somewhat reliable -- has>
>> shut down, due to ongoing and intractible problems with their Agfa
> >> processing machine.
> >>
> >
> >Sorry to hear about your lab. But with over three times the pixel count
> >of even the most expensive digicams, film and scanner still make more
> >sense to me.
>
>
> I've been using 35 mm film, plus film scanners, for several years now.
> I mostly agree with your conclusion, but even with 35 mm, there are
> problems getting good, consistent processing of C41 film.
I know what you mean. I quit shooting 645 print film because I couldn't
get decent machine prints locally. I have a place that does pretty good
custom prints, but they don't do machine prints, and they're high even
for proof sheets.
--
Carey L. Jones
Date: Thu, 31 May 2001
From: rafeb@channel1.com (Rafe B.)
Newsgroups:
rec.photo.equipment.medium-format,rec.photo.digital,rec.photo.equipment.misc,rec
.photo.marketplace.medium-format,rec.photo.misc
Subject: Re: Scanning 6x7 film?
daniel_a_robertson@my-deja.com (Daniel
A Robertson) wrote:
>I am an advanced amateur 35mm photographer, very seriously considering
>buying a 6x7 camera system.
>
>I have also been postponing buying a film scanner for about a year -
>almost bought a Microtek Artixscan 4000 (i.e. a Polaroid 4000), and am
>looking at the new Nikon stuff with interest.
>
>What REASONABLY-PRICED (a price fix to illustrate what I mean by
>"reasonably": I think the F5 is sold far too cheaply) scanner that
>will be able to handle 6x7s? Sure, 4x5 scanners are out there, but
>expensive, and I only want 33% of their areal coverage.
>
>Or should I just use a custom lab to send me scans?
>>
>Waiting for digital cameras to deliver comparable quality to the
>larger MF sizes is like waiting for Godot.
No, it's worse than that... I've been following this market
(hi-res, med-format scanners, and digital cameras) for a
few years now. Agonizing. I'm seriously considering
chucking film for something like a Fuji S1 or Canon D30.
Everyone's waiting to see how the new Nikon 8000ED pans out.
Any minute now... I've not seen any reviews of the Polaroid
120, either. Why is that?
Here are some models to consider. All of these would be found
"pre-owned" on places like eBay, or dealers that sell demos,
floor models, etc (eg., bitec.com):
1. Leafscan 45. An honest 2500 dpi on 6x7. $1800-$2500.
A huge machine, old design, but truly "pro" quality.
Rodenstock optics. Originally cost $10K-$15K, about
10 yrs ago. Service could be very expensive on this one.
It's large, heavy, and will cost a lot to ship.
2. Agfa 2500. A top-end flatbed scanner. An honest 2500 dpi.
$2500-$3000. Artix makes an equivalent model (same innards,)
but it is rare. The Agfa is about $4500 new. The Artix might
cost $500 less, purchased new.
3. Umax Powerlook 3000. $3000+. Top-end flatbed. But I'd
stay away from this one, for a few reasons. Plus, it's over-
priced for what you're getting (as a film scanner.)
4. Minolta Multi, Multi II. $800-$1500. Resolution is only
1100 dpi. Not quite what your film deserves. It'll scan
a smaller area (35 mm frame) at much higher resolution,
around 2500 dpi.
5. Nikon LS-4500. $1500. Only 1000 dpi.
6. Polaroid 45, 45i. The '45 is a mere 1000 dpi, no
better or worse res-wise than the Minolta and Nikon.
The 45i does 2000 dpi, and is still current (maybe) but
*very* expensive -- around $7000 retail. I've never
seen a 45i for sale on the used market.
7. Imacon Flextight. Ha! If you can find one, they should
sell for about $7K (pre-owned, in good condition.) I
saw one up for bid on eBay not long ago, but it never
sold. Owner was asking about $7500.
rafe b.
Date: Tue, 29 May 2001
From: "Amr Ibrahim" aibrahim1@houston.rr.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 120 Film Scanners
Glen:
I trust that you are in the States. If so, consider Acer 620UT (available
in USB and SCSI). It is for $99 in Office Max including the built in
slide unit. Its resolution is 600 DPI; which is not really fantastic, but
for the price is difficult to beat.
Best of luck
AI
"Glen Barrington" glenbarrington@home.com wrote
> Can anyone recommend a modestly priced (say, less than $800 US) scanner that
> will scan not only 120 film, but 35mm and aps film as well
Date: Fri, 20 Jul 2001
From: fulton@scantips.com (Wayne Fulton)
Newsgroups: comp.periphs.scanners,rec.photo.equipment.35mm
Subject: Re: digital archiving
joe@jodfs.com says...
>I wish transfer to CD-rom for archiving a as a continent way to show. I have
>about 1000 photos and I was wondering what is a good resolution and format
>to save them in. I am looking for a good balance between quality and size.
>(I have the canoscan 4000us and most software).
With CD-R disks costing only a few cents now, the difference in say 5
disks and 50 disks is a trivial factor cost-wise. Not much worse
organization-wise either, although you will need a cataloging system
either way. I think you want a second copy of all masters, just in case
one goes bad somehow, a scratch perhaps.
Scanning 1000 images is a lot of work to end up with the wrong result for
the future purpose. To avoid wasting all that work, you should work the
archiving problem backwards, from what you NEED as opposed to what is
convenient. It is hard to know all future purposes, but that does not mean
you shouldnt have a specific plan to work to. Perhaps that purpose is "as
good as possible for any unknown purpose", but you should know what you
are working for.
If the purpose for the images is only to view them on the screen in the
future, then a small screen size image (say 800x600 pixels) is sufficient.
If the purpose is only to view them, or possibly print them, but NOT to
have a good master for future editing purposes, then good quality JPG is
surely fine (if file size is actually important).
Such a small JPG image is probably fine for distributing to family
members, who will probably only spend a few seconds looking at each one on
the screen. But it is a poor plan for a "master" copy for any future
purpose.
If you may need to print them, how large? Is 6x4 inch snapshot size
likely more than enough? (most often, it is). Or do you need 11x17 inch
potential? Very different requirements, but the time to decide is now, not
then.
If you might want to edit those images in the future (if they are not to
be read-only in purpose) then JPG would be a poor choice, and TIF is
vastly better.
--
Wayne
fulton@scantips.com
http://www.scantips.com "A few scanning tips"
Date: Sat, 26 May 2001
From: Paul Butzi butzi@nwlink.com
Newsgroups:
rec.photo.equipment.medium-format,rec.photo.darkroom,rec.photo.equipment.large-format
Subject: Re: Sadly, I'm at a loss. Please advise...
"Michael J. Kravit"
kravit@bellsouth.net wrote:
>"Paul Stimac" pstimac@hotmail.com wrote
>> For a scanner, I'd get an epson 1640SU .......prints from
>> this scanner are as good as smaller prints from Tango Drum scans of
>> the same images.
>
>Is this right? Can you confirm this as fact, have you tested it to be sure?
It probably depends on the print size, and the image being scanned.
for my review of the 1640 SU photo scanner, I scanned several of Barry
Sherman's transparencies; Barry was kind enough to share drum scans
of those transparencies with me so that I could show side by side
comparisons. The scan I selected to show side by side drum scan versus
epson 1640su photo was done by www.nancyscans.com who coincidentally
appear to use a Heidelberg Tango drum scanner.
The comparison is not as good as I'd like; the two scans are not
color, density, or contrast matched. Nevertheless, you can see
a difference between resolution of the scans.
You can check it out at
http://www.butzi.net/reviews/epson1640suphoto.htm
I like that epson 1640SU a lot, and for the price it's a stellar
deal, but I don't think that at the bottom line it's going to replace
$70k drum scanners like the Heidelberg Tango.
But for $370, it's damn impressive. Maybe even stunningly
good. Far better than I expected, to be honest. Especially when
you consider that it cost about .5% of the Tango.
-Paul
--
http://www.butzi.net
Date: Sun, 27 May 2001
From: rafeb@channel1.com (Rafe B.)R
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Sadly, I'm at a loss. Please advise...
...
The 1640 certainly represents a rather singular price-performance
point, and it is one of the few scanners you can find for under $1000
that will do a "decent" job on medium-format transparencies.
But I'll be damned if its resolution is anywhere near the advertised
1600 dpi. Let's just put it this way... I can get a visibly sharper
8x10" print from a 35 mm negative scanned on a SprintScan Plus,
than I can from a 645 negative scanned on the 1640su. That should
not be possible, but it is.
OTOH, for $400, what can one expect? The Minolta Multi is around
$1500, pre-owned, and only has 1148 dpi (advertised spec.) on
medium-format film.
The best scans I've ever seen from medium-format transparencies
come from the Imacon Flextight scanner, and from the Leafscan 45.
rafe b.
Date: Thu, 14 Jun 2001
From: "Ralf Guminski" ralfgski@yahoo.com
Newsgroups:alt.comp.periphs.scanner,rec.photo.digital,rec.photo.equipment.medium-format
Subject: Re: scanning 120 negatives with Epson 1640SU
Jeannie,
If you locate some calibration software to utilise with the Q-60 slide you
can create a custom ICC profile for your scanner and eliminate one cause
of color caste. Each Q-60 has an exact color profile of its own which the
calibration software uses to "correct" the output from your scanner. Of
course this only affects colors output by the scanner's CCDs and not poor
colors in the film/photos being scanned or castes originating from your
monitor.
Ideally all components should be calibrated for color. One relatively
inexpensive product which can potentially do this is Colorific. At least
one person in these groups recommends it highly. I have part of the
software, the monitor calibration only, and it seems to have made a big
difference WRT correct colors right off the scan.
Cheers,
Ralf
....
From Rollei Mailing List;
Date: Mon, 23 Jul 2001
From: bob mcclelland bob@marscovista.com
Subject: Re: [Rollei] MF scanner resources...was...something else ;-)
Hi,
> I have heard that some people do like the results of the, I believe,
> 1640SU...which is available for under $200...
I took delivery of the 1640SU last week and must say I'm very impressed
indeed! The quality of the scans at 3600 on a roll film neg is really
good. Given that I haven't had it a week yet (so complete and stringent
testing has not occurred) I can wholeheartedly recommend one. In UK,
however, one will cost you o260 inc the tranny adapter.
regards,
Bob
Cornwall (U.K.)
www.marscovista.com
From Rollei Mailing List;
Date: Mon, 23 Jul 2001
From: Andrei.Calciu@hn.va.nec.com
Subject: [Rollei] Epson 1640SU scanner
Price in US is $168 ++$95 tranny adapter +$10 shipping. I found this at
http://www.technoweb.com
I never bought from this site so I cannot say if they are good or bad.
Andrei D. Calciu
From Rollei Mailing List;
Date: Mon, 23 Jul 2001
From: Andrei.Calciu@hn.va.nec.com
Subject: [Rollei] Epson scanner again
I just found a 400 dollar version of this same scanner. Here is the reason
why:
Box contents:
EPSON Perfection 1640SU Photo scanner,
CD-ROM:
EPSON TWAIN scan software,
Adobe Photoshop 5.0 LE,
Adobe PhotoDeluxe,
ScanSoft Textbridge Pro OCR;
ArcSoft PhotoPrinter,
ArcSoft Panorama Maker,
ArcSoft PhotoMontage,
EPSON Smart Panel,
USB cable and EPSON's Scanner Basics Guide,
Tranparency Unit
Andrei D. Calciu
From: "Leonard Evens" <len@math.northwestern.edu>
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Epson Perfection 1640 Scanner
Date: Sat, 18 Aug 2001
"Anthony M. Zipple"
<tzipple@rcn.com> wrote:
> Anyone scanning 645 negs with it? Opinions? Thanks
>
I've been using a 1640 Photo scanner for 6 x 7 and 6 x 6 negatives. The
6 X 6 negatives are usually cropped to 4:5 ratio.
I find it satisfactory but clearly not the results are not as good as
conventional
prints made with an enlarger. There is a subtle loss of sharpness which can
be
partially compensated for by use of an unsharp mask in Photoshop or Gimp.
But it takes quite a lot of work.
Of course, the advantages of working digitally are enormous, and if you
are willing to give a little on sharpness, you will probably find the scanner more than adequate, paarticularly for the price. But if you are
really
critical, nothing that sells for less than $2000 is likely to come close.
If you do decide to get this scanner, wait until the Fall when Epson's
newer versions of the scanner will be available. The optical
resolution has been raised from 1600 to 2400 ppi. But I don't think the
defects of the scanner have than much to do with nominal resolution, so
this may not make too much difference.
--
Leonard Evens len@math.northwestern.edu
Date: Tue, 4 Sep 2001
From: David Flanagan <dwf@caltech.edu>
To: rollei@mejac.palo-alto.ca.us
Subject: Re: [Rollei] Fixing scratches in Photoshop
Robert,
I think most people would agree that scanning from *either* a negative
or a slide (I won't step into that mess either) will produce results
superior to scanning from a print, for reasons that would include, but
not be limited to, the greater available density range and the
sharpness of the unenlarged image.
I may have mentioned before that I have enjoyed <italic>Adobe
Photoshop 6.0 for Photographers</italic> by Martin Evening (Focal
Press, 2001). It described a procedure for removing dust and
scratches by using that filter and the History brush, which I can
describe off-list if you are interested. Basically, that filter is
not intended for applying to the entire image, because it seems to
incorporate a combination of blurring and blending.
As a side note, that book also has the best description of color
spaces and color management that I have read, and makes it very clear
to a non-professional how color management and device calibration
should work.
Cheers,
Dave
Date: Wed, 10 Oct 2001
From: AJNECP@aol.com
Subject: Re: Scanning...
To: panorama-l@sci.monash.edu.au
Yes, Epson FlatBed Photo Scanners can handle longer negs or transparencies.
For Pro use the Expression series, but the late Perfection Photo are
outstanding, and can handle longer negs.
AJ
From: dilbertdroid2@aol.com (Dilbertdroid2)
Newsgroups: rec.photo.equipment.35mm
Date: 27 Oct 2001
Subject: Re: Hasselblad X-pan question.
The easy way would be the Nikon 8000 at about $3,500.00. I can get along on
stitching.>>>
You guys aren't listening. The easy way out is the HP S-20 which will easily
scan the whole Xpan frame in one pass at 24 dpi and is under $500. Does a
nice big 36mb scan on the Xpan frame.
From: "Sherman Dunnam" sherman-remove_this@dunnam.net>
Newsgroups: rec.photo.equipment.large-format
Subject: Re: flatbed for 4x5 and MF?
Date: Sat, 27 Oct 2001
"Stephen Ratzlaff" ratzlaff@ticnet.com> wrote...
> Please don't use mineral oil. There are several good solutions
> commercially available that perform this function.
>
What are the options other than mineral oil and why isn't mineral oil a good
choice? I have used it as described below with good result. Am I somehow
damaging my film or scanner?
There was a good article in CameraArts (from May this year I believe) on
using a flatbed Epson 1640 to scan 4x5 negs. The author recommended mineral
oil on the flatbed glass, lay the neg down carefully and remove any bubbles,
put mineral oil on the top of the neg again removing any bubbles and
covering the top with photo quality mylar.
I tried this using regular mineral oil and got the best scans I've ever
gotten with that scanner. Small scratches disappear and the overall quality
of the scan is higher. The drawback is the mess. I clean the negatives
with film cleaner which seems to do a fine job and the scanner with a soft
cotton cloth and glass cleaner.
Sherman
From: David Strip David.Strip@home.nospam.com>
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Epson 2450 & med format
Date: Tue, 15 Jan 2002
Leonard Evens wrote:
> "David Strip"
> David.Strip@home.nospam.com> wrote:
>
>
>>David Albrecht wrote:
>
> Where does one find a scan source to do such a test?
They may be acquired from
Applied Image Inc.
1653 E. Main St.
Rochester, NY 14609
call or write:
Paul Anderson
panderson@appliedimage.com
716 482-0300, x-214
ask for:
Scanner SFR and OECF targets
1 target = $60.00/each
5-9 targets = $50.00/each
10 or more targets = $34.40/each
These are simple 3 inch x 3inch targets.
>> The MTF drops below .5 at about 8 lp/mm, or 400 dpi.
>>
>
> 8 lp/mm should be adequate for high quality prints.
The context of the original question was scanning film. After paying all
that money for lenses that will resolve 40 lp/mm and still hold an MTF
of .5 or better, it seems that giving it all away on the scanner is
counterproductive.
In the end, they're your pictures. If *you* are satisfied with the scan,
then it's good enough.
Date: Fri, 31 Aug 2001
Subject: Re: [Rollei] Fixing scratches in Photoshop
From: Bob Shell bob@bobshell.com>
To: rollei@mejac.palo-alto.ca.us>
The easiest was to fix scratches is with an app called Kai's Photo Soap.
It has a tool that lets you drag a cursor along the scratch and it "pulls"
pixels from both sides to fill in the scratch. I haven't found anything
like it in any other app I've tested.
Bob
> From: "Philippe Tempel" ptempel@home.com>
> Date: Fri, 31 Aug 2001
> To: rollei@mejac.palo-alto.ca.us>
> Subject: [Rollei] Fixing scratches in Photoshop
>
> Can anyone point me in the direction of using Photoshop to
> fix the evil and ever present scratches and dust? They have
> a filter for it but I can't get any decent visible results from it.
> A pointer to a good book for RTFM is okay as well. BTW,
> I recently printed three of my photos on an old Epson Stylus
> Photo and like the output. Only thing is since the printer is
> 720 dpi I only get an image about as large as the 6x6cm neg
> itself (at 1600 ppi). So I had to resize them larger and got
> decent 6 inch prints. I tried to scan at 4800 ppi interpolated
> and gave up after waiting for over three hours for my scanner
> to have only 1/3rd of it done. :-(
>
Date: Sat, 22 Dec 2001
From: Adri de Groot adrijanus@juno.com>
Subject: Re: Epson 2450 Photo Scanner?
To: panorama-l@sci.monash.edu.au
Sorry, it's called:
Prazio
Anti-newston sprays
http://www.prazio.com
You can get it directly from them. They even have a sampler kit you can
buy.
DFStein@aol.com writes:
>
> adrijanus@juno.com writes:
>
> Otherwise
>
> use Praxio antinewton spray. >>
>
> Where can we obtain this?
From: "Leonard Evens" len@math.northwestern.edu>
Newsgroups: rec.photo.equipment.medium-format,comp.periphs.scanners,comp.graphics.apps.gimp
Subject: Sharpening with Epson flatbed scanners
Date: Mon, 24 Dec 2001
I've been scanning medium format negatives (6 x 7, 6 x 9, and 6 x 6)
using an Epson 1640 Photo scanner. I've been frustrated by the fact
that the scanner, while pretty good for the price, couldn't deliver the
detail and "photographic" quality I was used to in prints produced in the
darkroom. I've fiddled quite a bit with sharpening techniques, but I
seemed to reach a barrier. Fortunately there was a recent discussion in
some newsgroups about a technique which emphasizes sharpening edges,
and which is apparently well know by experienced hands.
With it, I've been able to produce something approaching what I want.
Indeed, I would say my 8 x 10 color prints look better than some I
produced years ago in the darkroom. (The darkroom still has an edge for
b/w.)
I wonder if there are other tricks I don't know about that might
be useful.
Here is how I use the technique (in the Gimp) for those unfamiliar with
the method.
I first apply a Gaussian blur to selected regions such as the sky in a
landcape.
I copy the image to the buffer, create a new channel, and paste the
copied image to the new channel, which will function as a mask.
I apply the edge detect filter to the grayscale channel mask. With the
gimp, I find a setting of between 2 and 4 works best. The larger the
setting, the narrower the edges.
I use either the levels command or the threshold command to clean up the
edge detected image.
I apply a Gaussian blur of 1 pixel to the result.
I copy the resulting mask back to a selection in in the original image.
It helps to toggle off displaying the selection. Sometimes I apply a
quick mask and fiddle a bit with the selection. (In Photoshop, one must
make sure to invert when copying back, but that is not necessary with the
Gimp.)
I then apply two successive unsharp mask sharpening with settings
in the Gimp of 1.8 radius and 50 percent in amount. (Gimp sharpen radii
don't correpsond to those in Photoshop.)
I finish by applying unsharp masks to selected areas which seem to need it
and sometimes I use a selective gaussian blur to remove some noise from
the image. Settings of about 3 and .3 seem to work best in the Gimp for
medium format.
Any suggestions on how to improve this technique even more would be
appreciated.
--
Leonard Evens len@math.northwestern.edu 847-491-5537
Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Can I scan MF film using a flatbed scanner?
From: Kyle Rhorer rhorerDELETE@hotmail.com>
Date: 28 Dec 2001
flexaret2@aol.com (FLEXARET2) wrote
> Try the closeout bargain flatbed - for under $79 -
> the Umax Astra 2200. It scans medium format negatives and
> transparencies really well.
Don't believe the hype. I received my Astra 2200 last week and have had
nothing but grief from it. Umax apparently hasn't come out from the rock
under which they live to realize that it's almost 2002. If you're using
Windows 2000, the Umax drivers won't work in USB mode without major
tweaking and hacking. Their VistaScan software is crap. If you manage
to get the scanner working, you will be very disappointed in your scans
of color negatives. Color transparencies and B/W negatives seem to work
fairly well, but you can forget about color negatives. VistaScan doesn't
know to remove the orange mask, and when I try to do it in PhotoShop I
find that there is no information left in the blue channel. The best
results I've been able to obtain look like a crappy Polaroid shot (the
kind you get from the $29 camera and 600 film).
Every piece of film I've scanned has come out looking grainy, even at
600ppi. In color scans, there is a lot of color noise. Newton rings are
a problem, although I've overcome that one. All in all, this experience
is a reminder that you generally get what you pay for -- there's a reason
film scanners are $2000-$5000!
If someone is obtaining satisfactory results with color negatives on an
Astra 2200 connected to a SCSI port under Windows 2000, I welcome a step-
by-step explanation of the process. Keep in mind I've spent the last
week scouring the Google archives (both Usenet and Web) for answers and
have tried most of the techniques I was able to find. Right now, the
scanner doesn't even work under Windows because I tried to install
MagicScan (from the Umax UK site). Of course I switched from USB to SCSI
before doing this, because MagicScan doesn't support USB. No joy. I can
get scans under Linux, so I know the scanner is talking to the SCSI port,
but the quality of the scans obtained that way is about the same as what
I was getting from VistaScan.
From: "Brad Swanson" brad.swanson@sympatico.ca
Newsgroups: rec.photo.equipment.medium-format
Subject: HP xpa Slide and negative adapter tips for MF Scanning
Date: Sun, 17 Mar 2002
After several months of working with my flatbed scanner with slide adater I
have come up with a few tips that will substantially increase the quality of
the output of the scan. Keep in mind I am a hobbiest and usually only make
prints to 8x10 but have printed as large as 11x17 from scans of negatives
from my flatbed with impressive results.
The scanner I have tested this on is the HP Scanjet 5370c with xpa adapter,
and my prints were made from a epson stylus PHOTO 1280. Although I do this
mostly for a hobby, I am presently working on a collage diploma in art
photography, and I have gotten good enough quality from my prints that I
generally charge $15-$20 Canadian for reprints.
In order to get a quality scan from a medium format film however you must
first make up a useable template to fit the factory 5" square plate. I made
mine with a piece of black construction board(99cents at Wal Mart). I
followed the following setps:
1. cut one 5" square, two 51/2" squares.
2. find the center of on of the 51/2" squares and measure out 2 3/16" both
ways from center and use those marks to draw two parrallel lines. With an
exacto knife, cut out the center section so you have two spacer strips.
3. layout the larger square and glue the two spacers so they are parrallel
and even with the edges, leaving a 2 3/8" space in the middle for the film.
then center the 5" square over it all and glue it to the spacer strips.
4. after the glue dries, measure the center of the 5" square and layout the
lines with pencil for the size of opening you need (ex. 6x7, 6x6). Then cut
the opening out with a straight edge and knife, through all the layers.
Now you have a working template, I have found there is still a problem with
light loss or intrusion so I made spacers (3)for between the xpa and the
factory template. Simply measure and cut out a square the size of the xpa
and cut out the center being sure to leave enough of a opening that the
calibration stip is not covered and the negative is open between the scanner
and the light. Make up 3 and glue them together so little light can escape
or enter and lay them all out like a sandwich on the flatbed.
These spacers alone should substantially increase the quality of the scans
you get from a medium format film, but I have other tips for HP users also;
1. seal the corners of the main template by putting a paperweight on each
corner or make up a more rigid template so less light gets in or escapes.
2. never,never,never block the calibrtion strip or you will get hideous
coloured lines through you scan. If you have to, turn the template 90
degrees and correct it in your photo editing program after you scan it.
3. check www.hp.com often as new drivers are always being introduced.
4. try to get the colour and hue as close as possible to what you want
during the scan, not afterwards in photoshop.
5. transparencies are fickle, and every spec can be seen an an enlargement,
clean the glass on the scanner and blow off negative before every scan.
There you go, quality MF scanning for under $500 Canadian. If you can add
to this thread or have comments, I look forward to reading them on this
thread.
Reguards,
Brad Swanson
Date: Fri, 22 Mar 2002
From: Tom Just Olsen tjols@online.no
To: hasselblad@kelvin.net
Subject: [HUG] HASSELBLAD & IMACON
Here in Scandinavia the word is out that Imacon, making possibly the
best scanners in the World, has reduced prices with 30% to DKK 40.000
(US$ 4,400). In the US, I can see, they have launched a campaign
together with Hasselblad that is worth looking closer at:
http://www.hasselbladusa.com/
What you need to run this beauty with medium format 'raw materials' is a
Mac or a PC, with some enormous RAM capacity. Go for 1 Gigabyte as a
start...
Tom of Oslo
Date: Wed, 17 Apr 2002
From: "M. Denis Hill" denis@area360.com
To: panorama-l@sci.monash.edu.au
Subject: RE: Scanners to handle 6x17
You might subscribe to the mailing list on scanners at
http://www.leben.com/lists/scan/, or the filmscanners list you'll find at
http://www.halftone.co.uk/. I don't recall where to find the archives of
these lists, but I do recall this comparison.
M. Denis Hill
Qualified Panoramic Photographer
rec.photo.equipment.medium-format
From: nyphotoboy@aol.com (NYphotoboy)
Date: Sun Apr 28 2002
[1] Re: Scanners for MF
I've been looking for the same thing... a MF slide/negative scanner that
doesn't cost $8000!! So far it looks like the Epson 2450 is the hands-down
winner, and nothing comes close in that price range.
The problem is... it's out of stock EVERYWHERE. The best price I've seen is at
buy.com, for $352, and you can sign up for the email notification when it comes
back in stock.
Kerry
rec.photo.equipment.medium-format
From: ramarren@bayarea.net (Godfrey DiGiorgi)
[1] Re: Scanners for MF
Date: Sun Apr 28 2002
Another vote for the 2450. I've been using it for several months now and
am 100% satisfied with the performance. I have scanned and printed
everything from Minox to 616 format negatives: it's certainly at its best
with larger format negatives, but does a remarkably passable job even with
subminiature film.
Godfrey
From medium format mailing list:
Date: Sat, 06 Apr 2002
From: Stan McQueen stan@smcqueen.com
Subject: Re: [medium-format] Digest Number 301
you wrote:
>Fordo:
>
>That's a gorgeous shot and thanks for taking the time
>to upload and share it. May I ask a question? How did
>you scan the neg?? I bought an Epson 2450 and I just
>can't seem to get the hang of scanning negs...they
>come out looking terrible and I can't seem to get the
>settings right. Or maybe the scanner just isn't up to
>the task. It wasn't expensive...it's a flat bed with
>transparency adapter. I see other posts on the net by
>people who get great results with this scanner...your
>input would be appreciated...
>
>frank weir
The 2450 is a great scanner for scanning medium format and 4x5 negatives
and transparencies. It's probably just about the best scanner available in
its price range (IMHO). Don't expect the images to come out of the scanner
looking great with any scanner, though; that's what Photoshop is for. I use
the Epson twain scanner software to crop the scanning area and set the
resolution and bit depth, but I don't try to adjust levels or curves there.
When the scan comes in, it's usually a little flat and muddy--that's
considered good, because it means I have captured everything that is in the
image without clipping any shadows or highlights. Setting black and white
points and touching up the curves in Photoshop then usually yields great
results. You might also try Vuescan (www.hamrick.com). I use that on my
35mm film scanner (a Polaroid SprintScan 4000). You might want to sign up
for one or both of the scanner mailing lists to get more info about
scanning slides and negatives. You can subscribe to Tony Sleep's
filmscanner list by sending mail to to listserver@halftone.co.uk, with
'subscribe filmscanners' as the subject or in the body. You can subscribe
to Mitch Leben's list by going to http://www.leben.com/lists and clicking
on the appropriate links.
Stan
Photography by Stan McQueen
http://www.smcqueen.com
From: pstimac@hotmail.com (Paul Stimac)
Newsgroups:
rec.photo.equipment.medium-format,rec.photo.darkroom,rec.photo.equipment.large-format
Subject: Re: Sadly, I'm at a loss. Please advise...
Date: 26 May 2001
For a scanner, I'd get an epson 1640SU for $370. The scans are amazing
relative to the cost of the equipment. There so good that I'd be
worried if I were competing scanner manufacturer. There's an article
in the new Camera Arts mag about people getting almost drum quality
scans from it. I just bought one and am impressed. Smaller prints from
this scanner are as good as smaller prints from Tango Drum scans of
the same images. For a printer to use, Epson 1160's are selling for
around $250 if you can find anyone that still has one. For about $620
you can have a darn good system for making great prints up to 13x44.
Of course the bigger the negative the better the prints, so I'd stick
with tthe mamiya and bender or Pressman, which ever is more
comfortable for you to use.
Good Luck.
...
From: rafeb@channel1.com (Rafe B.)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Sadly, I'm at a loss. Please advise...
Date: Sat, 26 May 2001
...(quotes above posting)
With all due respect, I say, hogwash. I own the 1640su, which
I use for scanning 645 negatives. I also own a Polaroid
Sprintscan Plus for scanning 35 mm.
IMHO, the Epson doesn't deliver anywhere near its rated 1600
dpi resolution.
I've posted some comparisoon scans at:
http://www.channel1.com/users/rafeb/scanner_test2.htm
You can see that the Sprintscan at 1350 dpi beats the pants
off the 1640 SU at its claimed 1600 dpi.
At the URL above there are links to scans from a
Minolta Scan Multi, Agfa 2500, and Leafscan 45.
BTW, I do agree about Epson 1160s' I have two of them.
Of course, we're all waiting to see how the new Nikon
8000 ED is going to stack up. At this point in time, the
venerable Leafscan is still the cream of the crop, in
the scans I've seen.
Maybe that Polaroid 120 can give it a run for the money.
rafe b.
From: ramarren@bayarea.net (Godfrey DiGiorgi)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Scanners for MF
Date: Mon, 29 Apr 2002
See the article at
http://www.normankoren.com/Tutorials/Epson_flatbeds.html for ideas on
how to optimize your focus. The 2450 optimum focus point is some small
distance above the glass, not directly in contact with the glass, so you
are best to experiment with shimming up the film carrier to bring the
emulsion to the sharpest focus point.
Godfrey
pleyenaa@direct.ca (Peter Leyenaar) wrote:
> Hi, I have an 2450 also And I am pleased with especially for the
> price.
> I have been wondering and experimenting with how to position the
> slides on the scanner, do you use the supplied film holder ?
> I have tried placing the slides directly on the glass of the scanner
> and placing a slide mount glass(Gepe) on top to keep it flat.
> Does anyone have any better ideas?
> Any workable suggestions would be welcomed,(no ,I don't want to put
> gel on my scanner ;-)
> Thanks Peter
From russian camera mailing list:
Date: Tue, 30 Apr 2002
From: "kowenphoto" kowenphoto@yahoo.com
Subject: Moskva 5 and the Epson 2450
I have one post for now from my Moskva 5 and a newly acquired Epson
2450. The shadow detail from this scanner is pretty good; the lens
on my camera tested sharp-yay!!! And the rangefinder also is spot
on. If you want to see the post, go to:
http://www.geocities.com/kowenphoto/
That will get you to my opening page (it will open somewhat slow) but
hit the link atop the page to the Epson 2450-it is on page 6. The
film is Fuji Reala, metered at the rated ASA 100. This is just one
example, more will follow. If you notice a greenish cast on all but
the last example, it can be corrected with the color cast removal
tool in the prescan software. The last one is corrected. Thanks-
kowen
From: "Roger N. Clark" rnclark@qwest.net
Newsgroups: rec.photo.equipment.large-format
Subject: Re: enlarger versus digital
Date: Mon, 18 Mar 2002
Concerning Brians' comments below, I completely agree with.
Are you a comfortable advanced computer user? If so,
then digital may well be the way to go.
I've had my own darkroom for 15+ years, and used
a bathroom for the 15 years before that doing traditional
developing and printing. I still do film, and occasionally
develop it myself, but I scan the film and print digitally.
In my opinion, digital touch-ups, color correction, contrast
stretching, etc are vastly superior than traditional wet chemical
and enlarging. And sharper results too!
For up to ~12x18 inch color prints, ink jet
prints are superior to any traditional custom color pro
lab print I have ever had done. And lightjet 5000
cibachrome prints from large format film is awesome!
Again far better than any traditional prints from a custom
pro lab.
I'll never look back. But it is more expensive, much more.
The computer, scanner, printer costs more than a traditional
darkroom. But Once set up, 12x18 inch ink-jetprints I bet are
cheaper, I think, than a color print. But those lightjet
prints can be significantly more (perhaps double)
that of traditional.
Someone in this thread mentioned pixelation in ~12x18 inch
prints. I think they were referring to small digital cameras,
not scanned 4x5 film. Drum scanned 4x5 velvia can produce
beautiful jaw dropping sharp 40x50 inch prints.
Here are my links:
A simple scanner comparison for 4x5 (you can do very well
these days with a $500 scanner, and for < 16x20 inch prints
can't tell the difference from expensive drum scanned images):
http://www.users.qwest.net/~rnclark/hp7400-drum_compare.htm
The digital print advantage over traditional prints:
http://www.users.qwest.net/~rnclark/digital_advantage.htm
Scanning detail for 4x5, other:
http://www.users.qwest.net/~rnclark/scandetail.htm
If you hate computers, 4x5 and traditional enlarging
can produce amazing, and beautiful prints--after all
people have done it well for decades. Producing
a fine print is an art, and a skill to be proud of.
Roger Clark
Photos at:
http://www.users.qwest.net/~rnclark
Brian Ellis wrote:
> Of course this is the kind of question that starts many bitter arguments.
> For me digital has cost far more than my traditional darkroom and will
> continue to cost more as the ever improving equipment makes existing things
> obsolete. I'm not heavily into digital but in the brief time I've dabbled in
> it I've bought two scanners, two printers (and getting ready for a third),
> added lots of memory, bought Photoshop 4, bought Photoshop 6, bought lots of
> inks at a cost of about $50 a set, bought lots of papers, etc. etc. In that
> same time I've added nothing to the traditional darkroom I started about 8
> years ago and have no plans to add anything in the future. My only costs
> have been chemicals and papers which aren't insignificant but certainly
> don't equal the cost of all the digital stuff. . So I'd be surprised if i
> turns out that you're correct in saying the cost is about the same.
>
> While many would disagree, I'd say that if you're doing color go with
> digital - you can do so much more with digital color than you can in a
> traditional color darkroom. If you do mostly black and white, you could go
> either way. I do exclusively black and white and the route I've chosen has
> been to dabble in digital and try to learn it while continuing with my
> traditional darkroom for the moment at least. Of course the main question,
> assuming you aren't doing this for a living, is which do you think you would
> enjoy the most? Some people hate sitting in front of a computer, some people
> hat messing with chemicals.
>
> "John Gunn" John_Gunn@btinternet.com wrote
> > hi everyone,
> > I'm just about to move up to 5x4, the camera side of
> > things I'm quite happy about. The question is which way to go on the
> > printing side of things, am I going to be better off buying a 5X4 enlarger
> > and producing prints in the darkroom, or would it be better to go down the
> > digital road and scan the negs and print them out on a good printer. I
> can't
> > see I would want to print above about 16"x12". Costs seem similiar for
> both
> > options, so has anyone any thoughts on the quality, lifespan of images or
> > any other useful comments on which system might be best
> >
> > thanks in advance
> >
> > john gunn
From: ramarren@bayarea.net (Godfrey DiGiorgi)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanning 6x17
Date: Wed, 08 May 2002
The Epson 2450 can handle scanning up to a 4x9 inch transparency at 2400
ppi. 6x17cm is no problem. You'll need to make a negative carrier to hold
the film and raise it up to the optimum focus point, shouldn't be much
problem with a bit of card stock to work with.
While this is not a scanner in the class of the Flextite, it can give
awfully good quality scans when used effectively and costs only $400.
Godfrey
"neil" neil@n2.com wrote:
> Hi everyone,
> I asked this question in another newsgroup and got no responses - hope I can
> get something here...
>
> I'm considering buying a flextite photo scanner primarily for private use
> (scanning 6 x 17cm transparancies). Due to the high cost of this equipment
> I was wondering:-
>
> 1. Has anyone used something better for the job for an equivelant or
> cheaper price?
> 2. A dealer recommended a Nikon Coolscan 8000- taking two scans of the
> neg then joining them - any thoughts? Has anyone tried this?
> 3. Would there be people interested in a service offering high quality
> scans of medium format trannies at reasonable prices?
>
> Thanks,
> Neil
From: bhilton665@aol.comedy (Bill Hilton)
Newsgroups: rec.photo.equipment.medium-format
Date: 08 May 2002
Subject: Re: scanning 6x17
>From: "neil" neil@n2.com
>I'm considering buying a flextite photo scanner primarily for private use
>(scanning 6 x 17cm transparancies). Due to the high cost of this equipment
>I was wondering:-
>
>1. Has anyone used something better for the job for an equivelant or
>cheaper price?
You mean the Imacon Flextight? Most reviewers and users feel it's probably the
best desktop scanner available for scans up to 6x17 cm, but expensive.
>2. A dealer recommended a Nikon Coolscan 8000- taking two scans of the
>neg then joining them - any thoughts? Has anyone tried this?
I have the 8000 (but don't need to scan film larger than 6x7 cm). In theory
this might work, lying to the scanner and saying you have two 6x9 cm images
(6x9 is the limit for the 8000) and then using a panorama or other stitching
program to joing them. In practice I don't really know how well this would
work, it depends on getting a near perfect exposure match from the scanner at
the junction of the two slices I'd guess.
For sure I'd want to see this demonstrated before plunking down $3,000 for the
Nikon :) Ask the dealer for a demo ...
Michael Reichmann is a well-known pro who has very good reviews of both the
Imacon (which he uses) and the Nikon 8000 (which he reviewed and compared to
the Imacon last fall). You can find both reviews on his site
http://www.luminous-landscape.com/
I know he uses the Imacon for 6x17 cm so after reading the reviews you might
email him and ask what he thinks of the "8000 and stitching" idea. He may know
of someone who tried this.
>3. Would there be people interested in a service offering high quality
>scans of medium format trannies at reasonable prices?
When the Imacon first came out people were spamming the medium-format newsgroup
with offers of 3200 dpi medium format scans for as low as $25. I don't think
they got much business, at least not from the newsgroups. Most people are
afraid to ship out their original film and if they decide to do this for scans
would rather deal with a larger professional scanning service even if it costs
a bit more.
Good luck.
Bill
From: ramarren@bayarea.net (Godfrey DiGiorgi)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanning 6x17
Date: Wed, 08 May 2002
The 2450 is fixed focus. They set the optimum focus
point to be some small distance above the glass platen then use a small
aperture to allow DoF to accommodate the "satisfactory sharpness" zone.
This makes allowance for originals which are not absolutely flat (not in
direct contact with the glass).
When you're scanning transparencies in medium format and larger size
range, the focus is satisfactory for most purposes. When you go to smaller
bits of film, focus becomes increasingly at issue to obtain the sharpest
possible results.
The article which clued me in to optimizing focus for very small format
film scanning is at http://www.normankoren.com/Tutorials/Epson_flatbeds.html.
I did focus testing by obtaining a very sharp test negative of a clearly definable
pattern (a Minox negative on Technical Pan film of newsprint at 3'
distance, focussed critically). Technical Pan grain is too fine to be
imaged clearly at 2400 ppi and cannot in any way drive the imaging process
into grain aliasing; the Minox lens is capable of up to 200 line pair per
mm resolution as well, so I knew this was going to give me a good test
target. I made a rigid holder for the Minox negative which was .010" thick
(a small piece of sheet plastic) and taped the Minox negative to it
securely so that it was as flat as possible. That put the central point of
the negative backing .010 mm above the glass. I made multiple scans, each
time adding a paper shim .003" inch thick to the bottom of the holder
(.003" just happened to be the thickness of two sheets of a PostIt note
pad at the glued end of the sheets). I did this for the range .010 to
.055" and found that .043" gave me the sharpest rendition of the image on
the scan.
It's relatively simple to check and determine focus this way. You can go
empirical and just keep adding shims in increments to see if it gets sharp
or blurry, stop when it starts to get more blurry.
Godfrey Sylvester fish@sylvester.fsworld.co.uk wrote:
> > The Epson 2450 can handle scanning up to a 4x9 inch transparency at 2400
> > ppi. 6x17cm is no problem. You'll need to make a negative carrier to hold
> > the film and raise it up to the optimum focus point, shouldn't be much
> > problem with a bit of card stock to work with.
>
> Is there another focus adjustment available.
> For instance, raising the scanner glass.
>
> How does Epson 'focus' the 2450? Or do they just assemble them?
> Seems a bit sloppy to offer 2400x4800 when maximum sharpness isn't
> available. I mean, who wants a hi-res soft image? You can get that through
> interpolation.
From: Jefro jefro@mcn.org
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanner with full page transparency capacity
Date: Sat, 18 May 2002
You are correct, the 1640 is not in the same league. The 1640's
transparency adaptor is 4x9, max optical resolution is 1600dpi. Dmax is
low enough that you wouldn't want to use it for archival work. The 2450
has a higher reported Dmax and optical resolution, but still the same
4x9 transparency lid. I get great scans from mine, but alas it cannot
be used conveniently for contact sheets.
Then again, the 1640 is only about $200 and the 2450 is only $350, while
the Expression is somewhere around $1100. Try them out in a store
before you buy, if you can, or hunt the web for someone who scans using
an Expression. You can find a great writeup on the 2450 here:
http://www.virtualtraveller.org/epson2450.htm
Robert Feinman wrote:
> The epson 1600 series ( I think the latest is 1680) scan full 8 x 11
> at 1600 dpi. There is a lower priced model, called 1640, I believe
> that is not in the same league. The "pro" model comes with a
> transparency adaptor. There are slide holders for 15 mounted 35mm
> slides as well as strips of 35mm, 120 and 4x5.
> You can select all the images in the scanning software and it will
> scan them in at once.
> You can tell you're looking at the right model if the price is about
> $1000+).
> I have the prior model, it has enough resolution for medium and large
> format slides and will make ok 8x10 enlargements of 35mm slides. If
> you want to go bigger you really need a 2000 - 4000 dpi scanner.
From: Sylvester fish@sylvester.fsworld.co.uk
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 6x7
Date: Sat, 11 May 2002
> Want to scan my b/w negs 6x7. Does somebody use Epson Perfection 1650 Photo?
> Is it good? Thanks, Jean Strinckx
>
I picked up an interesting link concerning MF and the 2450,
from the MF Scanning Service in Southern CaliforniaY thread.
http://www.photographical.net/minolta_pro.html
IYve cancelled my 2450 order until IYve checked out some sums.
Sylvester.
Date: Sat, 29 Jun 2002
From: "Brian" patience@e-mailanywhere.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: MF Scanner survival ?
"Justin Case" justicase@nospam.org wrote ...
> "Brian"
> patience@e-mailanywhere.com wrote:
> The Epson 2450 scans at 2400 dpi, full hardware resolution, no
> interpolation. If you're getting sub-2000 dpi scans its operator
> error, not the scanner.
I wish I was able to share your trust in manufacturer's specifications. The
following links did little to convince me.
1. a $40,000(!) Scitex at 2220dpi vs. the $400 Epson 2450 at "2400" dpi:
http://www.acmeenterprises.com/photography/equipment/pages/epson_2450.php
2. a sub-$3,000 Nikon Coolscan at 2400dpi vs. the Espon at "2400" dpi
http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=0026Vq
3. http://www.crystalcanyons.net/TechNotes/ScannerResolution.htm
4. http://www.geocities.com/aanticulturee/escantest.htm
You might also compare multiple tests from the same source. Pop Photo
reports the resolution of both the 4,000dpi Polaroid Sprintscan 120 and the
Nikon Coolscan 4000 ED as 60 lp/mm. Perhaps I imagined it (I only glanced at
the magazine while waiting to catch a plane), but I seem to vaguely remember
that PoP reported resolution from the Epson 2450 as around 20 lp/mm. . . .
I've enjoyed using my Epson 2450 to scan 6x7 since I received it late last
year. I just don't think that it resolves what the product literature says.
From: Leonard Evens len@math.northwestern.edu
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: MF Scanner survival ?
Date: Sat, 29 Jun 2002
Subhash Tiwari wrote:
> Recently bought an old Mamiya 645 to explore MF, but have no scanner.
> Initial trial with the Epson 2450 disappointing.
Have you looked at the various examples which some of us have posted
showing examples of what the 2450 can do. See
groups.yahoo.com/group/SeePhoto/files/Godfrey/scaneg/epson2450resolution.htm
www.math.northwestern.edu/~len/photos/pages/e2450.html
for two sets of examples.
My experience is that at normal viewing distances, I can produce prints
from medium format scans which are virtually indistinguishable from
similar prints I produced in the darkroom. Many of my scans are from 6
x 6 negatives but are cropped to the 6 x 4.5 format.
You would undoubtedly do a bit better with a $3000 film scanner, but the
difference would only be apparent under very close examination.
Just what did you find disappointing about the Epson 2450?
> Have considered Canon
> D2400UF , but hesitate because of reports.( Love my Canon 4000 for 35mm.)
> Don't want to pay $ 2.5-3K for the high end Nikon/Minolta/Imacon scanners.
>
> Said to someone that will wait for better/ less expensive /Great quality MF
> scanners (All this stuff gets better/cheaper with time- right ?). Was told
> waiting might be a bad idea since Film/scanners will all go the way of the
> Dodo, and be replaced by digital cameras/backs etc.
>
> So, should one wait ??
>
> Subhash
--
Leonard Evens len@math.northwestern.edu
Date: Wed, 17 Jul 2002
From: Leonard Evens len@math.northwestern.edu
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Advice on scanning 4 x 5 and other matters?
Ralph Barker wrote:
> Actually, Adobe did have a port of Photoshop for Solaris a few years
> back - around Photoshop 3 if memory serves. I suspect they eliminated
> UNIX support based on the lack of volume sales. It's just too costly to
> support a particular OS unless the volume is there.
Maybe that is only the case if you are a large corporation. Ed Hamrick
doesn't seem to have any problem providing an excellent scanning program
that runs under Windows, the Mac OS, and Linux. It supports
practically every scanner, and he sells a license allowing unlimited
upgrades for $40. But you can't expect a large corporation like Adobe
to be as efficient as one good programmer. :-)
--
Leonard Evens len@math.northwestern.edu
Date: 5 Aug 2002
From: "Jim Phelps" Jim.Phelps@MI.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: K64 - would it sell?
...
One thing I've experienced is a large amount of color crossover printing
most E-6 films on Ilfochrome (never fully understood what the problem was
until I read _Post Exposure_ by Ctein). I never have a problem with PKR.
Although I've reduced a great deal of crossover using Provia 100F, it is not
eliminated. No mater what I do, I cannot get all of the colors right all of
the time. If you want to see what I mean, take a couple of good shots of a
MacBeth color chart and print it on Ilfochrome. The orange square is way
off, and several of the others aren't that close either. When I do this as
a test for my style of printing, I aim to get the primary colors right, but
don't try to print skin with it... This is also the reason why I love
Kodachrome, and have for years (since my first experience with it comparing
it against old E-4 films). I guess I could always stop using Ilfochrome and
start using R-3000, but like with Kodachrome, it's going to be a divorce and
not a parting of the ways :(
Anyone have a combination of slide film that prints well on Ilfochrome
(except Kodachrome of course)?
from hasselblad mailing list:
Date: Sun, 14 Jul 2002
From: Darrell Jennings darrell_jennings@yahoo.com
Subject: Re: [HUG] Update on Nikon 8000ED scanner
Fritz, I was told by Epson that the 2450 only goes to
1200 DPI optically. Also I found that was the case
when I was driving it with Vuescan.
--- fritz olenberger olenberger@cox.net wrote:
> The Epson handles 2400 optically. I agree, the
> extra resolution of the
> Nikon probably isn't what is making the difference.
> Lenses, software,
> glassless carrier probably are what makes the
> difference. By the way, does
> someone know which of these scanners (Epson, Nikon,
> Minolta, Poloroid)
> generate three true colors per pixel (tri-linear
> arrays?) versus using a
> single array with color interpolation?
> -Fritz
> ----- Original Message -----
> From: "Darrell Jennings"
> darrell_jennings@yahoo.com
> To:
> Sent: Sunday, July 14, 2002
> Subject: Re: [HUG] Update on Nikon 8000ED scanner
>
>
> > I was only considering prints, and compared to the
> > Epson at any resolution, 1200, 2400,4800 since the
> > Epson only handles up to 1200 optically, the
> > additional resolution didn't seem to buy much on
> it.
> >
> > I'm not convinced that the 4000DPI is really doing
> it
> > in the Nikon either since my printer won't handle
> > anything like 4000 DPI. I think it has to do with
> the
> > improved lighting (three CCD's).
from hasselblad mailing list:
Date: Sun, 14 Jul 2002
From: fritz olenberger olenberger@cox.net
Subject: Re: [HUG] Update on Nikon 8000ED scanner
Printing: The term DPI (dots per inch) should be reserved for printers.
Those little pico-liter ink dots are not pixels. Set your epson printer to
1440 dpi and then forget you ever heard about dots. Tests have been done on
the Epson Photo 1280 printing at 1440 dpi and at 2880 dpi, and visually,
there is practically no difference.
You want to size the image in Photoshop to print at 260 to 300 PIXELS per
inch (ppi). Even the best Chromira silver halide prints go with this
resolution.
Scanning: The difference between scanning at 4000 pixels per inch and
scanning at 2400 pixels per inch is that you end up with more pixels
(assumming the same scanning area), so you can print a bigger image at the
same resolution of 300ppi.
-Fritz
from hasselblad mailing list:
Date: Sun, 14 Jul 2002
From: fritz olenberger olenberger@cox.net
Subject: Re: [HUG] Update on Nikon 8000ED scanner
I would get a second opinion from Epson. He may be correct, but that is a
suprise to me. The Epson web site advertises the 2450 as having 2400x4800
HARDWARE resolution. I always use the lower number and consider that
optical resolution. I think the 4800 number has something to do with the
stepping motor. The interpolated resolution goes much higher.
-Fitz
from hasselblad mailing list:
Date: Sun, 14 Jul 2002
From: Godfrey DiGiorgi ramarren@bayarea.net
Subject: Re: [HUG] Update on Nikon 8000ED scanner
Yes, 4800 ppi is achieved by halving the stepper motor.
The 2450 has a hardware optical resolution of 2400 ppi, not 1200 ppi.
This is borne out by my comparison with a Minolta Scan Dual II (2820
ppi). The resolution between the two scanners is much closer than it
would be if the Epson were only capable of scanning at 1200 ppi
optically.
Godfrey
From hasselblad mailing list:
Date: Sat, 10 Aug 2002
From: Beau beauh@rbnet.com
Subject: Re: [HUG] Who's a good "back" mechanic?? plus 2450 scanner tip at bottom!
> Dick Werner, the author of the manuals, is a great guy and will tell you
where to get the parts and help you with anything you don't understand.
Thanks Jim! Dick was kind enough to write me and I've already pummeled him
with a couple of newbie questions. He *is* a smart guy!
When I went to HB's site, they directed me to many dealers, most of whom I'm
sure hire excellent technicians. But I was wondering tho' - are these
networks of HB dealers the only way to repair equipment or could one have HB
itself repair things?
It's not that I feel that a dealer can't do an excellent job - I'm just
curious how HB has set it up. Thanks again!
beau :)
PS - so as not to veer too far down another scanner thread - I thought that
all you folks with the new Epson 2450 MF scanner (which HB is "giving" away
with their X-pan in a recent promotion) might find this interesting: The
scanner is not very adept at scanning 6X6's because of a poorly designed,
plastic transparency adapter. I saw in some British photo publication a
*generic* transparency adapter (holds negs/chromes while scanning) for 6X6's
made by Umax touted for *any* flatbed scanner. Anyone who's cussed the
2450's holder like I have might be interested - I haven't tried it yet, but
I think it's only around $25.00. Just FYI - check Umax's site.
[Ed. note: see Diffraction pages for more on limitations...]
From: "Roger N. Clark" rnclark@qwest.net
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Woes in scanning 4 x 5 images.
Date: Sat, 10 Aug 2002
Doug Dolde wrote:
> Roger,
>
> I have a drum scanner (Screen DT-S1030ai) that will go 5200 dpi max. Do you
> think it's worth going that high on a 4x5 transparency?
>
I'm jealous ;-)
Well, why not try? I think it depends on your image.
If you look at my page:
How much to sample to record "all the detail?" at
http://www.clarkvision.com/imagedetail/sampling1.html
I conclude that scanning at 2 to 3x the Dawes resolution limit
of the lens (assuming the lens is the limit, not film grain)
gets all the detail.
At f/64 this works out to about 2200 to 3300 ppi. If the image was
done at f/32 and still be diffraction limited, then double these
ppi's. Of course, you need fine grained film, like Fuji Velvia
to reach these levels.
Dawes limit for other f/stops is at:
http://www.clarkvision.com/imagedetail/scandetail.html#diffraction
But scanning at this level is only slightly more detail than
scanning at say half the ppi. Depending on your application,
it may not be worth 4x the file size! For me it is.
Roger
From: aaron@post-modern.net (Aaron van de Sande)
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Woes in scanning 4 x 5 images.
Date: 10 Aug 2002
There seems to be alot of confusion here over gross dpi and halftoning
resolution. There are plenty of resources on the net that describe the
difference.
http://www.naples.net/partners/kwikkopy/scanning.htm
--Aaron
> the best way to go. For example, with the Epsons, while resolution is
> stated to be 1440 or 2880, depending on the model, etc., the Epson
> printer driver for Windoze appears to discard input image data beyond
> about 300 DPI, and then fiddles with what's left to produce its output
> resolution. If using an outside printing service with commercial,
From minolta MF mailing list:
Date: Sun, 14 Jul 2002
From: David Kilpatrick iconmags@btconnect.com
Subject: Re: Re: advice on scanners
Re Nikon and Minolta - the biggest difference between the scanners is in
the handling of colour negs and the potential for really good colour
management.
Nikon scanners are well respected for natural colours from negs, and
used by most pressmen for this reason alone. Minolta scans from
negatives are not in same league (the only other scanner make which
manages a decent, natural film-like colour from negs is Polaroid and
they, like Nikon, use film-specific tables - Minolta does not do so).
The Minolta scanners generally capture a colour gamut which looks
'wrong' out of the box - frequently too pinkish, lacking especially in
yellow and green, from colour slides. There are two solutions to this -
first, use proper colour management. Generally, Mac users running
Colorsync 2.5 or later with the latest Minolta scanner drivers get
stunningly good colour from reversal scans, and even the neg quality is
decent since there is a separate profile now for negative operation.
These profiles are applied internally, and set up in the scanner
dialogue. They can not be applied after doing a raw scan, in Photoshop;
the result is wierd in the extreme, I have of course tried!
Ed Hamrick (Texas, not England!) has produced a very decent universal
scanner program in Vuescan. I am using Mac OS X on my main production
machine at the moment, and I am obliged to use Vuescan to keep working
in a purely native OS X (Unix based) environment. His latest revision
still will not handle rollfilm scans on the Multi/Multi II (all you get
is the central 24mm wide strip) but without using any colour profiles,
it includes some basic corrections which create better colour negative
scans by a long way than Minolta's driver, and almost as good slide
scans. However, if you use Vuescan you lose ICE etc. The main benefit of
Vuescan is that I can have a flatbed scanner and two slide scanners
attached, and Vuescan will operate all three, just choose from the
window - also it recognised almost any scanner you care to stick on the
machine. The list is huge.
BUT Vuescan does not match, in any way, the optimum performance of a
scanner like the Multi Pro or the Elite II when CORRECTLY used with
colour management - i.e. set your colour space (the first window) to the
same set up as your monitor, and your file export colour space to the
same space as your Photoshop working space (i.e. sRGB or Adobe RGB, the
second window of the two). Many people simply don't know which to enter,
or do not check this option, and don't use the supplied colour
management, or use PC systems where colour management has not been
properly set up (Mac systems almost make it impossible to work without
setting up a colour workflow, and OS X has it so deeeply embedded that
it works with ALL programs even those with no provision for control).
If you do not use colour management, and do not make sensible decisions
about when to use auto exposure for each image, and when not to, and how
you use the extensive image controls - you won't see what a Minolta
scanner can do. Also, if you owned either the original Dual or the Scan
Speed or original Elite, Vuescan will probably give you better colour
anyway. The ancient QuickScan 35 despite a very limited density range
gives surprisingly fast, pleasant and accurate colours. The Elite II,
Dual II, and the revised Multi II upgrade (alters the firmware control)
for the Multi all give greatly improved colour and make best use of the
sophisticated and accurate colour management which comes with the scanners.
Ignore this software, and you'll get pinkish-looking, muddy shadowed,
lackluste scans; use it and you get vibrant, tonally accurate results.
Hamrick's Vuescan has not improved my productivity or accuracy, but
keeps us going until October when Max OS X drivers are expected for the
Scan Multi II which is our main production workhorse.
It is possible to get results on a PC which fully equal those on a Mac,
and there is no substantial difference in performance between the two
platforms today; in 1986 when we bought our first Mac there was of
course so much difference there was no choice. However, the mindset of
industry Mac users who have been brought into colour management
gradually since around 1990, predisposes them to hunt down 'settings'
religiously and to have all the programs, drivers etc fully
synchronised. Most of the complaints against Minolta scanner colour,
etc, are from PC users using uncalibrated (or 9300 G2.2!) systems and
ignoring Windows CMM or ColorSync, and thus getting serious mismatches
between scan colour space (or lack of) and image handling program
settings. So before you listen to advice from professionals, especially
from press photographers, be sure to find out just exactly what they
KNOW about scanning apart from being happy with what they see from their
own kit!
PS: anyone who views any images on any kind of laptop, LCD, TFT screen
etc is unqualified to comment on anything. I know, I just bought a
lovely LCD 18", and that's the end of being able to see colour
accurately, ever. I am quite sad. My calibrated Diamondtron 19 inch
screens were up to soft proof standards; this LCD thing will probably
save my eyesight and not fry the brain cells with emissions, but one
thing it can NOT do is to show photographic colour properly. But I like
it so much I am learning to trust 'adjusted' judgment.
David Kilpatrick
Minolta Club
http://www.freelancephotographer.co.uk
from hasselblad mailing list:
Date: Sun, 14 Jul 2002
From: Moreno Polloni mp@dccnet.com
Subject: Re: [HUG] Update on Nikon 8000ED scanner
> Darrell said "significantly" better, you said "noticeably" better. Those
> mean two different things to me.
> The reason I asked is that I had a friend
> scan a 6x6 negative on his Epson 2450 at 2400 dpi and I cut the same
> negative down to 35mm size and scanned it on my Polaroid SS4000 at 4000
> dpi. While the Polaroid scan did show slightly better detail on the screen
> I wouldn't call it significantly better and, in my test, you have to look
> closely to tell the difference in a 11" x 14" print from my Epson 1270
> printer.
The Epson scans are pretty good on their own. From a MF or 4x5 neg, the
images are very smooth and look sharp. However, the Nikon scans show not
only more detail in terms of sharpness, but the tonal contrast is more
pronounced, there's more shadow detail, less noise, and the colours are
cleaner, with subtle details coming through that are lost on the Epson.
It's always hard to qualify the differences, as they may mean different
things to different people. They may be "significantly" better on some
images, and just "noticeably" better on others. Given the price differences
between the scanners, you would expect the Nikon to be better, and it is.
The Epson is not as good, but for what it costs, it's a great bargain.
Date: Mon, 12 Aug 2002
From: Godfrey DiGiorgi ramarren@bayarea.net
To: hasselblad@kelvin.net
Subject: Re: [HUG] ... 2450 scanner notes ...
> Please see:
> http://www.normankoren.com/Tutorials/Epson_flatbeds.html
I've seen the article before and done my own experiments. Remember that
he was doing his work with 35mm film, at least as far as I can
discover, not 6x6. A couple of things I found:
- Finding the optimum focus point becomes more critical as you move
down in format size ... Small format originals give you fewer pixels to
work with no matter what you do, magnification to a useful image size
will be greater, so sensor DoF is effectively much slimmer.
- The 35mm film carrier supplied with my unit was further off from the
ideal focus point than the 6x6 format carrier is. Seems to indicate to
me that perhaps Epson has medium format and above in mind for this
scanner as its primary target.
- Most reflective scanning is performed best at resolutions between 150
and 1200 ppi, making the sensor's DoF effectively much greater.
I built a couple of small carriers for Minox, half frame 35 and Tessina
negatives to optimize focus for them, but honestly it cannot really
compete with the Minolta Scan Dual II for best quality from small
formats, both for reason of the Minolta's focusing optical system and
higher ppi capability. I use the Epson 2450 for 645 and 6x6 almost
exclusively now, which, you see from the example I posted, it does very
well with the default hardware. At least to the limit of its price tag
... the 10x step required for comparable or better dedicated film
scanner is bigger than I can justify or afford at the moment.
>> [903SWC photos]
>>
> Very good! Be sure to give us a link!! Best regards!!
I'll include the HUG in this week's PAW announcement. :-)
Godfrey
From: bhilton665@aol.comedy (Bill Hilton)
Newsgroups: rec.photo.equipment.35mm
Date: 21 Sep 2002
Subject: Re: Fillm scanner for medium format?
>From: "Randy Roy" randyr5@comcast.net
>Which film scanners will do medium format as well as 35mm?
Nikon 8000ED and Polaroid Sprintscan 120 do both film sizes at 4,000 dpi.
Minolta Multi Pro scans 35 mm at 4,800 dpi and MF at 3,200 dpi.
Imacon Flextight does MF at 3,200 dpi.
Each has its plusses and minuses, and its fans. None are cheap.
I have the Nikon and would recommend it highly.
Bill
From: ralf@free-photons.de (Ralf R. Radermacher)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Fillm scanner for medium format?
Date: Tue, 8 Oct 2002
John Eyles jge@cs.unc.edu wrote:
> There's a very strong
> rumor that they're following it up with a similar model, at
> a not too much higher price, that has 3200dpi (instead of
> 2400dpi) resolution.
Rumour? It's called the GT-9800 and it's on Epson Japan's website.
Ralf
...who bought a 2450, a mere two weeks ago.
--
Ralf R. Radermacher - DL9KCG - K"ln/Cologne, Germany
NEW URL!!! private homepage: http://www.fotoralf.de
Date: Sat, 19 Oct 2002
From: Christopher Williams leicachris@worldnet.att.net
To: hasselblad@kelvin.net
Subject: [HUG] $179.00 4x5 Microtek scanner
Anyone seen the new Microtek 5900 flatbed scanner in use? At $179.00 with a
adapter for up to 4x5 slides and negs, this may be a good proof scanner.
4800x2400 resultion, 48 bit. No info on Dmax though.
Chris Williams
New Orleans
Date: Thu, 24 Oct 2002
From: Godfrey DiGiorgi ramarren@bayarea.net
To: hasselblad@kelvin.net
Subject: Re: [HUG] Scanning Xpan 35mm Negatives
Yes, the 2450 will handle this very easily. I don't think the Epson
TWAIN software will do it, but I drive this scanner with VueScan
http://www.hamrick.com/ which allows any arbitary size negative that
fits within the 4x9" transparency scanner's space.
Godfrey
Steve Baker wrote:
> Can anyone guide me as to whether the Epson 2450 or something similarly
> priced will scan 35mm negatives for the Xpan? Also-any USA pro labs
> out
...
Date: Thu, 24 Oct 2002
From: Manu Schnetzler marsu@earthling.net
To: hasselblad@kelvin.net
Subject: Re: [HUG] Scanning Xpan 35mm Negatives
Actually, Epson TWAIN doesn't have any problem with arbitrary size. I've
scanned xpan negs with it with good results (such as
http://www.schnetzler.com/PAW/week_39.html).
The only problem I see is a bit of vignetting but that's a due to a lack
of center filter on the camera, not to the scanner.
Manu
Date: Thu, 24 Oct 2002
From: Leonard Evens len@math.northwestern.edu
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Epson 3200 dpi MF scanner out in Japan
Jeff wrote:
> "Bart van der Wolf" bvdwolf@nospam.nl wrote
>>85MB would equate to an uncropped nominal 6x6cm at 8-bits/color and
> 2400ppi.
>>Bart
> So this means that a 6x9 neg would yield a 130MB file.
> Is 8-bits/color sufficient for decent quality printing?
That depends on your definition of "decent quality". I've been making
inkjet color prints from 8 bit per channel files using Epson printers
for a couple of years now. I can't distinguish them from wet darkroom
color prints I made years ago in my darkroom. 8 bits per channel should
be beyond the ability of the human visual system to distinguish colors.
The main reason for 16 bits per channel is so that you have room for
manipulating the image without opening up gaps. If you increase the
contrast in a photoeditor, you may separate values too much if you are
working with 8 bit color. But I do that all the time, and I can't see
any degradation in color.
But perhaps I can't see subtleties that would drive you crazy.
P.S. With any kind of decent lossless compression, those files can be
cut down in size significantly when stored on disk. A modern high
speed CPU with a lot of ram should be able to handle 130 Mb files
without too much trouble. I have a 1.4 GHz Athlon with 1 Gb of memory
and I regularly process images of size 6600 x 8250 = 54.5 Megapizels at
8 bit color depth. That corresponds to a file size of 163 Mb
uncompressed. I use the Gimp under Linux. Things are a bit slow and
sometimes the photoeditor crashes if I try some really complicated
selection---the same thing happens with Photoshop 6 under W98---but it
is tolerable.
> Jeff.
--
Leonard Evens len@math.northwestern.edu
Date: Fri, 25 Oct 2002
From: "Gabriel Regalbuto" gabe@regalbuto.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Epson 3200 dpi MF scanner out in Japan
"I. N. Cognito" nomail@nospam.net wrote...
> "David J. Littleboy" davidjl@gol.com wrote:
> >FWIW, I found that the 2450 was quite soft. 1800 dpi Kodak ProPhotoCD scans
> >were significantly sharper and contrastier.
> I've heard others say this as well. I really wonder if the sample
> variation of this scanner is that wide?
I read a review of the 2450 which illustrated that the focus of different
units varied considerably. The author had shimmed his carrier and was
getting considerably sharper scans. I have recently gotten one of these
scanners. While for the most part satisfied (larglely due to the roughly
$2450 price difference between this and the Nikon), I am about to start
testing for my scanners optimum focus.
Date: Thu, 24 Oct 2002
From: Gordon Moat moat@attglobal.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Epson 3200 dpi MF scanner out in Japan
"David J. Littleboy" wrote:
> FWIW, I believe that compared to digital cameras, these raw scan files only
> have 30% the information D60 files have. So the 2450 would yield 12.6MP
> files from 6x9, and the 3250 might yield 22.6MP files.
A bit different effect is happening here. A flat bed scanner captures less
detailed information than a film scanner, and a drum scanner captures even
more. You could have exactly the same file size from each, and find more
detail, better colour information (and Histogram), and greater bit depth from
the drum scanner. This is very different than mega pixels, which are not a good
comparison.
> Whether the conversion factor is 25%, 30%, or 36% is a matter for further
> study.
>
> > Is 8-bits/color sufficient for decent quality printing?
>
> 8 bits is fine for printing on an inkjet. You need 16 bits if you are doing
> radical stretching of shadow areas.
>
> I wish people wouldn't talk in MB; it's really quite confusing. If you talk
> in MP + bit depth it's much easier to understand what you've got.
Perhaps as a comparison to marketing terminology used to sell digital cameras,
but even there all megapixels are not created equally. The use of MBs has been
used for longer than the somewhat current interest in direct digital cameras.
It is more useful for working professionals, since it gives somewhat of an
indication of manipulation ability in PhotoShop, or image placement limitations
in Quark. Even then, a file from a drum scan of exactly the same dimensions as
a film scan, will often have better detail in shadows and highlights. Bit depth
and Dmax are often overstated in specifications for scanners and digital
cameras, so YMMV.
> FWIW, I found that the 2450 was quite soft. 1800 dpi Kodak ProPhotoCD scans
> were significantly sharper and contrastier. My 2450, at least, is much
> softer and less contrasty on a pixel-per-pixel basis than any dedicated film
> scanner.
Not necessarily a bad thing, especially for portraits. I still use Kodak Photo
CD choices for some projects. They are still fairly good, cost effective, and
great for saving some time.
> . . . But those ridiculously overestimate the amount of information in those
> scans. My experience was that the 21MP I was getting from the 2450 with 645
> was somewhat better than the 5MP I was getting from the 5MP F707, but not as
> good as the 6MP D60. (IMHO, the D60 is worlds better than the F707.)
There is so much going on here. It would be tough to directly compare, but it
would probably be very enlightening to see a drum scan of a MF film, and
compare that. Comparisons on a computer monitor are often difficult to give an
indication of differences, though comparing Histograms may help a bit more.
> So if 645 + 2450 was a 7.5 MP image (which prints to a very nice 8x10), then
> 645 + 3250 might be a 13.3 MP image. These are images that will print nicely
> at 300 dpi. But these still are lower quality images than D60 images, which
> print nicely at 240 or even 200 dpi.
This would often have more to do with printing technology, paper choices, and
printing inks. There are so many different ways to print something, and there
are ways to optimize any file to better match a printing method. It could
easily be a situation that the default settings of an inkjet may be a better
match for some digital cameras, though the dot gain of any inkjet limits the
final image quality more than the information contained in the file.
> What I do with 8000ED images is downsample by a (linear) factor of .707,
> exactly halving the pixel count, in a sharpen-downsample-sharpen sequence.
> The result is an image that looks quite good printed at 330 dpi at 13 x 18.
> With the 2450, downsampling by .707 isn't enough and leaves a still soft
> image. Downsampling by .5 is a bit too much, 0.6 seems better. Optimal
> downsampling should result in an image that sharpens well in Photoshop at
> 100%, 1 pixel radius. Raw scans require 200% and a 2-pixel radius or so.
Any interpolation can introduce errors. This is a situation that requires
practice, and experience. Once you have matched settings for one printing
system, you often find that they do not work well for another printing system.
After you have enough things printed, you get more of a feel for what works,
what compromises to make, and how to quickly get the results you want.
Practice, practice, practice . . . and talk to your print shop.
> David J. Littleboy
> Tokyo, Japan
AGFA has a great series of inexpensive books about printing technology. Most of
them cover digital issues, image files, colour issues, and scanning technology.
There are no simple answers, and few direct comparisons, but the more
information you learn, the better your decisions will become.
Best of luck to you, and may the light be with you.
Ciao!
Gordon Moat
Alliance Graphique Studio
http://www.allgstudio.com/gallery.html
From: "Doug Dolde" doug.dolde@gte.net
Newsgroups: comp.periphs.scanners,rec.photo.equipment.medium-format
Subject: Re: Medium-Format Transparency Scanners?
Date: Wed, 20 Nov 2002
The Epson 2450 is a real bargain at $399 list. It has film holders as well.
No good for 35mm but for medium format and 4x5 (especially) it isn't bad at
all. I had an Imacon Photo which I sold. It's surely better but also $5K.
The Imacon pulls out much more shadow detail and scans at 3200 dpi compared
to the Epson's 2400. I actually like the Epson's color rendition better
however. It's a bit more saturated or vivid. That may not mean more
accurate but it looks good.
It's definately good enough for proof scans then you can get your prize
photos drum scanned.
Epson is soon phasing out the 2450 for a new 3200 dpi scanner. If you can
wait it would be the one to get.
"earl" earls@bcpl.net wrote ...
> Does anyone know of any decent (new) medium-format transparency
> scanners for under $1000? Thanks, Earl
From: Leonard Evens len@math.northwestern.edu
Newsgroups: comp.periphs.scanners,rec.photo.equipment.medium-format
Subject: Re: Medium-Format Transparency Scanners?
Date: Wed, 20 Nov 2002
earl wrote:
> Does anyone know of any decent (new) medium-format transparency
> scanners for under $1000? Thanks, Earl
I've found my Epson 2450 produces very good results. See
math.northwestern.edu/~len/photos/pages/e2450.html
for some examples.
The 2450 costs about $400, but it has essentially been discontinued. It
has been replaced in Japan with a similar scanner with an optical
scanning frequency of 3200 ppi instead of 2400 ppi. The newer scanner
should be available in the US in a few months, so you may want to wait.
On the other hand, you may be able to find a 2450 marked down some
time in the near future, and it may very well be entirely adequate for
your needs. It should be noted that even at 2400 ppi, scanning a 6 x 7
negative or transparency yields a very large file. So at 3200 ppi, the
files would be even larger. You need a fast computer with a lot of
memory to be able to handle such files in a photoeditor. My 1.4 GHz, 1
Gb ram, computer is just barely adequate.
--
Leonard Evens len@math.northwestern.edu
From: Leonard Evens len@math.northwestern.edu
Newsgroups:comp.periphs.scanners,rec.photo.equipment.medium-format
Subject: Re: Medium-Format Transparency Scanners?
Date: Sat, 23 Nov 2002
...
> Is that a Mac you're using?
> The reason I ask is because I scanned a 4 by 5 tranny on my D2400Uf at
> 2400 dpi and got a huge file, 285 megs or so. I have only 192 megs
> memory (Win98) and I had no problem handling it. Photoshop took a bit
> longer manipulating it, but that's to be expected. It wasn't that bad,
> really.
I have a PC and I use the Gimp under Linux and sometimes Photoshop 6
under W98. I can read the 2400 ppi image into the photoeditor with no
problem, and I can do a few operations. But anything really complex
takes too long and on occasion causes the program to crash---either the
Gimp or Photoshop. I have an older computer with 192 Mb, and even on
scans from 6 x 7 negatives, I have problems. So I think it depends on
what you are trying to do. I regularly do sharpening as follows. I
copy the image to a channel. I then run an edge detect filter on the
copy, increase the contrast with the levels tool, blur it slightly, copy
it back to the original image, and apply an unsharp mask filter. The
last step involves a very complex selection. Generally that is where I
have the biggest problems, when the selection is very complex. Also,
when the history mechanism involves a lot of old images or image
segments. Simple changes using the curves tool are easier to accomplish.
--
Leonard Evens len@math.northwestern.edu
From: "Todd" nashtn212@spamaol.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Newbie going from 35mm to medium-format
Date: Thu, 21 Nov 2002
It's funny you should write that when you did. Just yesterday (probably as
you were posting), I received my new Epson 2450 and was making my first
scans with it. I suppose it's the best sub-$500 scanner out there for film,
but I have to admit I was a bit disappointed. I had all this sharpness and
detail in my Velvia slides shot in 6x4.5 and 6x9 that was just completely
lost in the scans. It was frustrating. I guess I'm off to the pro lab to
have some drum scans made.
...
Date: Thu, 02 Jan 2003
From: MIKE GRACE amazing50@hotmail.com
To: panorama-l@sci.monash.edu.au
Subject: Re: Homemade film holder for a Noblex 120?
I built a holder from cardboard and glue to hold strips of 120 film. Works
well, just copy the size and thickness of the suplied holders. Make sure to
cut out the top notch as the scanner looks for this when in auto mode to
frame your negative. When you get one that works it is handy to build a
duplicate so that you can be setting up a neg while an other is being
scanned. Speeds things up greatly.
Good luck,
Mike Grace
Kitchener, On
Canada
From: David Kilpatrick iconmags2@btconnect.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: National Geographic Rates Film 400% better than latest Nikon D1X Digital
Date: Mon, 6 Jan 2003
Bill Tuthill wrote:
> Per Inge Oestmoen pioe@coldsiberia.org wrote:
>
>>VT wrote:
>>
>>>But one runs into aliasing and other loss problems there too
>>>(note the Nikon CoolScan 4000D tested at only about 60lp/mm when
>>>theoretically at 4000dpi it should have done closer to 79lp/mm......)
>>
>>Nikon LS-4000 only manages 60lp/mm? That is very bad. Do you have any
>>source for this claim? Not that I doubt it, but I would like to
>>document the shortcomings of the LS-4000.
It should only achieve 60 lppm (line PAIRS per millimetre for those who
don't grasp that this definition changed around 1966 when Geoffrey
Crawley was publishing his work on MTF - before that people used to
quote LINES PER MILLIMETRE and hence the figures given for early Leica
lenses are unfeasibly high, and should be halved to match modern lppm
readings).
It has 160 LINES (not LINE PAIRS) per millimetre scan density, deriving
this from 4000 lpi (lines per inch - lines, not line pairs). This equals
80 line pairs, which would be a 100 per cent contrast theoretical
resolution. 60 lppm at a normal film contrast level sounds like a
reasonable factor for light scatter, physical depth of emulsion,
diffraction limiting etc.
So that's not really a surprisingly low figure, just to be expected.
David
From: Paul Chefurka paul@chefurka.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: National Geographic Rates Film 400% better than latest Nikon D1X Digital
Date: Mon, 06 Jan 2003
Bill Tuthill ca_creekin@yahoo.com wrote:
>Per Inge Oestmoen pioe@coldsiberia.org wrote:
>>
>> VT wrote:
>>> But one runs into aliasing and other loss problems there too
>>> (note the Nikon CoolScan 4000D tested at only about 60lp/mm when
>>> theoretically at 4000dpi it should have done closer to 79lp/mm......)
>>
>> Nikon LS-4000 only manages 60lp/mm? That is very bad. Do you have any
>> source for this claim? Not that I doubt it, but I would like to
>> document the shortcomings of the LS-4000.
>
>The LS-4000 result is better than the Polaroid 4000, but not as good
>for high resolution as the Canon FS-4000, according to this webpage
>and related reviews on the imaging-resource website:
>
> http://www.imaging-resource.com/PRODS/LS4K/L40PICS.HTM
The 60 lp/mm figure (actually 61.something if memory serves) was actually
measured by Pop Photo when they did a set of scanner tests last year. This
is a real-life test scanning a slide of a lab-grade test target (not sure
which one, probably USAF 1951). I don't think there are many desktop
scanners out there that will do significantly better, as the resolution is
typically limited by noise.
60 lp/mm is well above the threshold required for large sharp prints (at
least up to A3 or 13x19), and is well above the typical resolution for most
film/lens/photographer systems, which seem to top out around 40 lp/mm. Raw
resolution is simply not an issue with the LS4000 or either of the other
4000 ppi contenders.
FYI, a "sharp" print usually exhibits a resolution of 4 to 6 lp/mm on the
paper. Given a real resolution of 60 lp/mm for the LS4000, you can enlarge
a scan 10 to 15 times and have it appear "sharp". At an enlargement of 12x
(A3) you'll have 5 lp/mm on the print, at a digital resolution of about 320
ppi. This is more than acceptable.
The web page referenced above gives the resolution (measured on the USAF
target) of the Polaroid as 50 lp/mm, the Nikon as 57 lp/mm and the Canon as
64 - which they note is the highest resolution they've yet seen from a
desktop scanner.
They also note that on their own home-brewed target, which they claim is a
more sensitive test than USAF, the Polaroid scores 2000 lines per picture
height, the Nikon to a bit past 1800, and the Canon somewhere just short of
2000.
The moral of the story is that the relative rankings of these three
scanners will change depending on the test being performed. Overall, they
are so close in performance as to make any choice between them dependent on
factors other than raw optical horsepower. Having used both the Polaroid
and the Nikon, I can say I'm happy enough with the Nikon not to lose any
sleep over it.
Paul
From: David Kilpatrick iconmags2@btconnect.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: National Geographic Rates Film 400% better than latest Nikon D1X Digital
Date: Fri, 3 Jan 2003
Paul Repacholi wrote:
> David Kilpatrick iconmags2@btconnect.com writes:
>>In a recent Master Digital photographer issue, Jason Smalley took
>>similar (near identical) shots side by side with Velvia in an EOS-1
>>and on his Canon D60. He used a Nikon 4000 dpi scanner to create a
>>DPS image from the Velvia, and interpolated the D60 image up to
>>provide a similar (55mb) file. Both he left unsharpened and
>>unprocessed.
>
>
> If a `master' did this, I hope you got your money back.
>
> This is a *BAD* comparison of a D60 and the Nikon scanner. If you want
> to compare them this way, you will need an astronomical plate scanner
> or the like. 11,000 ppi, 16bit, and then do it 3 time, one each for
> RGB. Otherwise you are just wasting your time wanking. And NG will still
> will have no interest in paying extra for your digital shots.
>
This is a practical comparison of the actual working methods used by a
well published and respected natural history photographer. It's what he
uses - and how an 'astronomical plate' scanner would help I don't know.
I use - when occasion demands - one of our pair of Leafscan 45
machines, which are 5,000 dpi and 14 bit and three pass, with filters
changeable by us, lamps replaceable, and still after 12 years not
surpassed by desktop machines.
Guess what happens at 5000 dpi? You just see MORE grain MORE sharply on
Velvia. Especially if you open up the Apo Rodagon fully, select a focus
point, run focus, stop down half a stop, and then examine the image from
that exact focus point.
I've been reproducing 35mm since 1975 or so and I've worked with all
kinds of scanners (except the type you mention - I cut my photographic
teeth in an observatory darkroom, but that was in the days of 5 x 4
plates and I was only 14 and had to repair all the gear and make my own
carriers for films). I can remember how the Crosfield and Hell reps
howled down the possibility of our ever using our first desktop scanners
for repro (Hasselblad MacSIE and, er, Microtek flatbed!). We did
comparison tests. We beat what they were doing in the 1980s - our
results looked better than the scans we had been paying o50 each for.
And then we found that these guys had been making all those claims about
density, sharpness etc and never bothered to tell us the scanners
reproduced just 64 steps per printed colour! Sure, they could penetrate
the sort of d-max Kodachrome offered, and they had some good colour
correction tables - but when we saw what we used to call a 'bad tone
break' on a colour proof, it was because those lovely big drum scanners
were really incredibly crude even by 1988 desktop standards. They
finally got round to 256 steps per ink sometime in the MID 1990s and
some repro houses and printers still rely on gear which is not even at
that stage.
Frankly, all that matters in photography is the final print - the image.
My experience of all film scanners is that they are next to useless when
confronted with colour or b/w neg. It does not mean I have never made a
decent scan from negs, but regardless of look-up tables, special
software etc, I still reckon a flatbed scan from a well made print beats
a filmscan direct from a neg any day.
As for slides, I'm sorry, but I have to scan slides all the time and
there are days I could weep. Eighty percent of slides which reach me,
even from agencies, are filthy and damaged; many are not sharp and even
with an 8X Schneider lupe it's impossible to tell. We used once to stick
them in an enlarger we kept in the editorial office, and look at them on
a white baseboard. It's clear the photographers have never projected
them or printed them so they don't even know they are not sharp. I
actually get angry with many competition entrants for sending me
wonderful looking slides which turn out to be quite useless because of
fundamental flaws, when they are examined (or even when they are
scanned). They're lucky they don't get them back watermarked with tears!
Digital, on the other hand - with a decent index sheet, with a properly
written CD? The standard of most submissions is excellent, well above
the standard from film-using photographers, and highly professional.
The same goes from rollfilm; we get a tiny percentage of duff rollfilm
work, but most is usable. The problem lies with 35mm slides. Most prints
we get are at least sharp - by the time the image is printed, I guess
the photographer can see it well enough to judge. But I seriously,
honestly think many photographers are just looking at their slides by
eye and sending them in to publications.
This is reality, and reality as MOST publishers are experiencing it
(notwithstanding National Geographic's remarkable head-in-sand self
justification).
David
From: David Kilpatrick iconmags2@btconnect.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: National Geographic Rates Film 400% better than latest Nikon D1X Digital
Date: Sat, 4 Jan 2003
David Littlewood wrote:
Re unsharp slides sent by so many people to mags:
>
>
> How large are you proposing to print them? I suspect many
> not-quite-perfect slides which would fail in a 16x12 Ilfochrome would
> look perfectly good in a 4x3 inch repro in a magazine - but you have
> lots of expertise in publishing and I have none, so tell me if I am wrong.
>
We scan all slides at 2820 dpi - whatever the native max res of the
scanner happens to be. We only ever scan at perfect multiples of the
scan res, 50 per cent, 25 per cent etc so really that means scanning at
full res. We do this because scanner interpolation firmware is usually
rubbish and produces unpleasant aliasing or other artefacts. All
sizing-down is done in Photoshop, but these days I am just as likely to
leave all scans as 25mb files, and size them on the page, and let the
PDF creation software do the downsampling. Then I can go back and change
my layout freely.
Recently I have let some full pagers go to press, which were only truly
sharp enough for quarter page. In future I must not do so. The TFT
monitor gives a slightly false expectation of sharpness compared to the
old CRT screens, and I need to be more critical in rejecting poor slides.
Actually, an unsharp image looks bad at any size, because the reader
subconsciously adjusts to examine small repros more closely. So even a
quarter page or eighth page image needs to start from a very good original.
David
http://www.freelancephotographer.co.uk/
From: VT vtVincent@CLOTHESaccess4less.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: National Geographic Rates Film 400% better than latest Nikon D1X Digital
Date: Sat, 04 Jan 2003
David Kilpatrick and I are of roughly the same vintage - we both took
part in the long defunct (UK) "Photo News Weekly" weekly Assignment
competitiion over a 26 week period in the 1960's - David came 2nd, I
came 4th in the final (league) standings.
So one may assume we both are reasonably well experienced with film.
The old publishing criteria was a good 10"x8" print -
which these days would be scanned at 300dpi to result in
3000 x 2400 = 7.2Mp
But 10x8 prints from 35mm in the 1960's were probably technically not
as good as ones from today's films and papers.
Estimates from some that about 6 to 8Mp is good enough for full page
publication is probably based on the above scanned from print
requirement.
Digital has the advanatge that the image is in "source" form without
losses from printing and scanning.
Film images can avoid the print loss by being scanned directly -
But one runs into aliasing and other loss problems there too
(note the Nikon CoolScan 4000D tested at only about 60lp/mm when
theoretically at 4000dpi it should have done closer to 79lp/mm......)
Commercial Drum scanning can yield much higher quality over any
consumer desktop scanner (thus the limitation of some comparisons -
is actually due to the limit of the desktop film scanner)
Having said all that National Geographic is well reputed for their
stunning images and quality - and their requirement is 400ppi for
images -
I would say that NG do know a thing or two about image quality and are
very unlikely to be deluded or misled by myths and out of date
requirements.
NG thinks a 6Mp digicam of high quality like the Nikon D1X is only
adequate for 1/2 page reproduction.
Of course opinions and YMMV -
but then none of us is actually National Geographic
and 400ppi is their requirment.
--
Vincent
David Kilpatrick iconmags2@btconnect.com wrote:
>In a recent Master Digital photographer issue, Jason Smalley took
>similar (near identical) shots side by side with Velvia in an EOS-1 and
>on his Canon D60. He used a Nikon 4000 dpi scanner to create a DPS image
>from the Velvia, and interpolated the D60 image up to provide a similar
>(55mb) file. Both he left unsharpened and unprocessed.
>
>We took a half page (A5) landscape slice of each image, and applied
>normal USM (modest, since we are a 200lpi sheetfed publication with very
>high levels of detail).
>
>I promise you that after seeing the comparison, you would never use film
>again. There's no way the Velvia can match the quality of the digital file.
From: VT vtVincent@CLOTHESaccess4less.net
Newsgroups: rec.photo.digital,rec.photo.equipment.35mm
Subject: Print resolution was Re: National Geographic Rates Film 400% better than latest Nikon D1X
Date: Thu, 02 Jan 2003
Gordon Moat
wrote:
> National Geographic does use 200 lpi screening (information from their media kit),
>which by the rule of thumb common in pre-press preparation, would dictate
>400 ppi.
Thanks to Gordon for his informative post.
The publishing requirements for National Geographic is 400ppi - this simply means
full (standard US letter sized 11"x8.5") page needs 4400 x 3400 pixels = 15Mp
But NG is slightly smaller-
10"x8" (?) page needs 4000 x 3200 = 12.8Mp
A double page (16x10) spread would then need about 25.6Mp
However Gordon also has said the majority of the magazine industry
work with 300ppi - this translates to:
Full standard Letter size (11x8.5) needs 3300 x 2550 = 8.4Mp
standard Double letter size (17x11) would then need ~17Mp
As mentioned coincidentally
300ppi is about 6lp/mm on paper - the traditional film print quality standard.
and 400ppi is about 8lp/mm on paper - the Leica film print quality standard....
--
Vincent
remove CLOTHES for e-mail
http://UnknownVincent.cjb.net/
Date: Thu, 02 Jan 2003
From: Gordon Moat moat@attglobal.net
Newsgroups: rec.photo.digital,rec.photo.equipment.35mm
Subject: Re: Print resolution was Re: National Geographic Rates Film 400% better than latest Nikon D1X
is more image information required to account for trimming of the page. With good press
registration, it is possible to only need 1/8" extra over each bleed edge. If the image
was only to the edge of the paper, a slight variation in trimming would show a white
line at the edge, rather than have the colour image go off the edge of the page.
It is interesting that the Kodak full frame digital just meets the ability to deliver a
300 ppi image at double page (Tabloid) size with adequate bleed. Many publications are
slightly smaller than full Tabloid size, making this immediately usable.
Anyway, if you are scanning for publication, allow for any printing to an edge by
giving more room to trim the pages. If you can put more than 1/8", or get up to 1/4",
that should account for any trimming, registration, of alignment changes.
I am still not sure that the lp/mm works directly. A 400 ppi or 300 ppi image would be
processed through a RIP an output on an image setter, or press, at 2400 dpi or 2450 dpi
(depending on manufacturer). This 2400 dpi output is for each colour, with a different
angle, and overlap for each of the four colours. The alteration of the patterning can
provide for a continuous tone appearance. It would be simple to say that 2400 dpi
relates to 47 lp/mm (roughly), but that is not the entire situation. It is also no
possible to directly state that 150 lpi printing is 75 lp/in. or about 3 lp/mm, since
that would only be true for one colour printing, and even then is not entirely true,
since it is too much of a simplification.
Since common magazine printing is four colour (CMYK), on a 2400 dpi press, trying to
take a roughly 47 lp/mm number, and multiplying by 4 (for each colour) would give 188
lp/mm. While that sounds great, the factors that mess that number including dot gain,
colour saturation, and registration errors. That leaves only measuring to really tell,
though a look at some good photography books can reveal some amazingly detailed printed
photos. Perhaps we should just acknowledge that printing quality can be quite high.
Ciao!
Gordon Moat
Alliance Graphique Studio
http://www.allgstudio.com
VT wrote:
> Gordon Moat moat@attglobal.net
> wrote:
>
> > National Geographic does use 200 lpi screening (information from their media kit),
> >which by the rule of thumb common in pre-press preparation, would dictate
> >400 ppi.
>
> Thanks to Gordon for his informative post.
From: "Jeremy" jeremy@no-spam-thanks.com
Newsgroups: rec.photo.equipment.35mm,rec.photo.equipment.medium-format,rec.photo.darkroom
Subject: Re: Getting a good scan of a print
Date: Wed, 26 Feb 2003
snip]
PRINT (not a negative) to post on the web.
I have scanned hundreds of old family album shots, and I have tried various
techniques, but there was one particular tip that I received from someone on
the Paint Shop Pro Newsgroup that has proven to be INVALUABLE to me:
DO NOT DO ANY CORRECTION AT THE TIME OF SCANNING--DO ALL CORRECTION IN YOUR
EDITING SOFTWARE AFTER YOU HAVE SCANNED THE PHOTO.
I use an HP scanner. I set my resolution at 300 dpi, and I apply NO
sharpening. I set my Exposure Adjustment as follows:
Highlights=255 (the maximum)
Shadows=0 (the minimum
Midtones=2.2 (the default setting for midtones on this scanner)
The scan ususlly comes out dark & muddy, but the scanner has captured as
many pixels as it could, making it easy to not go into your photo editing
software and adjust brightness, contrast, color saturation, color balance
and sharpening. And, you WILL have to adjust each scan--they NEVER come out
looking right when you first scan them.
By applying no correction at the scanner level, you will always start out
from the same base when you scan, and you will achieve more consistent
results.
I recommend that you also check the www.scantips.com web site and especially
review the HISTOGRAM section, as it is essential that you learn how to
"adjust the points" to bring out all the highlight, midtone and shadow
information in your prints.
I also recommend the following site which details steps to take in improving
pictures. It is oriented toward users of Paint Shop Pro, but the principles
can be used with other image editing software as well:
Finally, I recommend that you consider switching to Paint Shop Pro,
especially if you are not a PhotoShop power user. The learning curve for
Photo Shop is much steeper. While it is better suited for professional
applications, I have never found Paint Shop Pro couldn't handle anything I
needed. Believe me, there are LOTS of features in PSP--many of which you
will never use.
I have found digitizing old family prints, and distributing them to a much
wider audience, to be a most gratifying project. It was well worth the many
hours spent learning the proper techniques. Truly a labor of love--at least
in my case it was. Good luck on your project.
From minolta mailing list:
Date: Fri, 28 Mar 2003
From: "Peter Blaise Monahon peterblaise@yahoo.com" peterblaise@yahoo.com
Subject: Re: OT + + + Printer Resolution + + +
> Date: Fri, 28 Mar 2003
> From: "mr800si" songsin.sungkhawaetai@roche.com
> Subject: OT + + + Printer Resolution + + +
>
> I'm considering to buy an Inkjet printer. However, one thing I still not
> clear while reading the specification sheet is printer resolution. I know
> that normally printer resolution will express in DPI, like 600 1200 or
> 2400. But in some printers, manufacturer put the resolution like 4800 x
> 1200 dpi. What is the different between the 1st and the 2nd number?
> I've tried to find the explanation in Internet but so far not successful. All
> of them only mentioned with only one DPI number.
>
> Thanks for all the answers.
>
> Cheers,
> Tom
--
One number is the movement of the printhead sideways, the other number
(usually the larger number) is the movement of the paper down. It's
cheaper to make a paper-moving motor that can move in 1/2 increments
than it is to make twice more print head nozzles of 1/2 size. This has
NOTHING TO DO WITH YOUR IMAGE/PRINT JOB RESOLUTION (I know I'm
yelling). The above printer probably can handle a 720 dpi print job
regardless of the number of nozzles. See the printer manufacturer or
buy it and try it, and try some program like Norton System Information
to see the printer device characteristics through software are different
than the glorious print head advertising claims - same for flatbed
scanners.
See http://www.ddisoftware.com/qimage/quality/ for help with this.
Also, we may all want to budget some money for an up-scalling program
that smooothly invents the missing pixels we might need if a standard
printer driver blow up s-p-r-e-a-d-s out the scan pixels too far for
your liking - check out:
Qimage at http://www.ddisoftware.com/qimage/quality/ and then
http://www.ddisoftware.com/qimage/
Lizard Tech (ne Altimira) Genuine Fractals at
http://www.lizardtech.com/solutions/photo/
Human Software's Xfile (and more) at
http://www.humansoftware.com/pages1200/XFile/HSxfil0.html
... and so many other resources - search for
[digital image enlargement software plugin photoshop] at
http://www.google.com/
... and have a blast with the 150 responses (hey, that's small - we
could review that in one night's reading - have our credit card in
hand!).
Note that Qimage and Xfile brag about using a sophisticated "Lanczos"
filter to resize and poor Adobe Photoshop only uses a simple ol' bicubic
filter which is ugly at the extremes ... hmm ... I've been using FREE
IrfanView to resize images for ages (instead of Photoshop) and IrfanView
uses a sophisticated "Lanczos" filter to resize, and did I mention that
IrfanView (for the PC) is FREE?
Free!
http://www.irfanview.com/
Check it out.
Other programs to schmooze an image are stand alone plugins (outside the
scanner) from folks like Applied Science Fiction's GEM, ROC and SHO,
Polaroid's scratch and dust filter and so on.
Having fun yet?
We thought our purchases were over, but we just open up a subscription
entitling us to buy more, and more, and more ... !!! Yikes! The gift
that keeps on taking!
Check out ASF Applied Science Fiction's supplemental image scmnoozing
programs at
http://argon.asf.com/asf/download.asp?pid=1000&tc;=9998
Polaroid dust and scratch filter FREE at
http://www.polaroid.com/service/software/poladsr/poladsr.html
--
Tom, this may seem like overkill for your apparently "simple" question,
I know, but why waste time? Let's get to the deep, deep research right
now!
Peter Blaise Monahon - peterblaise@yahoogroups.com -
http://www.peterblaisephotography.com/
------------------------------------------------------------------------
PS - Here are my current printer "driver" resolutions regardless of the
published print head or other on-the-glossy-box specifications:
HP LaserJet 6p - 600 dpi (black and white proofs)
HP DeskJet 952c - 30 dpi (color proofs)
Acrobat PDF - 96 dpi (hey, they only expect screen viewing, right?)
Acrobat Distiller - 1,200 dpi (this is for pre-press proofs and fpo for
placement only)
HP DeskJet 660c - 300 dpi (color proofs)
HP LaserJet 5L - 600 dpi - (text)
Apple LaserWriter II NTX - 300 dpi (doesn't everyone have this? - for
Postscript troubleshooting)
Envoy - 200 dpi (like faxing)
MS Imagesetter - 600 dpi (I leave it, MS installs it and I've stopped
fighting)
Symantec Fax - 204 x 196 dpi (gotta be different, don't they!)
That's just one machine.
... so I use IrfanView to "Lanczos" resize and sharpen (and save if
needed) and print an exact dpi to the destination. Having a Wintel
computer, I can use Norton System Information to quickly see the
internal software driver resolution dpi and then I can hit the target
without guessing.
------------------------------------------------------------------------
PPS - Why are we thinking that so many of our threads, like this one,
are OT Off Topic? I propose that if someone really has an OT post,
PLEASE simple include an off-group link to where you'd like to continue
the discussion, otherwise, it's de facto on topic! =8^o
From: "Rick Rieger" rrieger@wideopenwest.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 6x7 scanning
Date: Sun, 23 Mar 2003
I just got the Epson 3200. Great scanner, except the software installation
was a mess. Under Windows XP, an obscure MS patch is needed. Without it,
the Smart Panel software installation will present you with a bunch of
install errors. See www.newsoftinc.com who was the author of the Smart
Panel software. They tell you what to do.
This is the second time I've had fits with Epson software. The last time
was with a 2200 printer...locked up XP so it would no longer boot. Epson
claimed it was XPs fault. Finally found a workaround.
Rick R.
"Smitty" jschmidt@uslink.net wrote
> Is there such a thing as a reasonably price scanner for 6x7 negatives?
> Smitty
From: fcarello@tiscalinet.it (Fernando)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 6x7 scanning
Date: 21 Mar 2003
"UrbanVoyeur" nospam@nospam.urbanvoyeur.com wrote
> Get a used Minolta Scan Multi I or II. The II might even be available new at
> a heavily discounted price, since it has been surpassed by the Pro. Either
> Scan Multi will produce results that are vastly superior to any flat bed.
Please note, that Multi and Multi-II (not Multi Pro) are only capable
of 1410 "true" dpi when scanning MF as (not so clearly!) stated by
Minolta itself; while they can go 2820 dpi for up to 35mm.
The Multi-II goes 2820 *interpolated* when scanning MF, while Multi
can do that only with a firmware/driver update (or it will only do
non-interpolated 1410 dpi).
While I do agree that they do a better job than a flatbed anyway (even
than those new "3200 dpi" ones: the point here is lens performance,
not CCD performance), I think it was fair to point that out.
In a similar way, the Multi Pro can do 3200 dpi in MF scanning, but
4800 in 35mm scanning: 4800 dpi in MF are interpolated (from 3200).
Regards,
Fernando
From: fcarello@tiscalinet.it (Fernando)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 645 versus 35mm????????????
Date: 6 Apr 2003
I'm responding to the original poster now.
The reason why your scanned 645 looks grainier than scanned 35mm is
because you asked for the same *final image* size.
So, your lab performed the 645 scan at a lower digital resolution than
35mm.
So, aliasing occurred: the optical system of a scanner is designed in
accord to its maximum sampling resolution; its MTF curves will have a
low-pass behaviour to act like an anti-aliasing filter, for the
nominal sampling res.
Scanning at lower rates, the optical system cannot perform its
anti-aliasing function, so the lens captures details that the sensor
cannot properly sample: this leads to aliasing, that appears as grain
in your scanned image.
Contrary to most common sense, even with fine-grained film (Provia
100F) you can have grain aliasing.
This would not be, if an auto-adjusting anti-aliasing filter could be
adopted, to adjust optical frequency response to actual selected
sampling resolution, but scan makers just don't take this path.
Well, next time I suggest you ask for maximum sampling resolution for
both scans, then you'll eventually resize (bicubic or bilinear, it
doesn't matter that much) in Photoshop to the final image you need.
This way, you'll notice the 645 beating up badly the 35mm. ;)
Regards,
From camera makers mailing list:
Date: Thu, 15 May 2003
To: cameramakers@rosebud.opusis.com
From: Robert Mueller r.mueller@fz-juelich.de
Subject: Re: [Cameramakers] lens in front of a scanner
Thank you for sharing your experience!
This is not a new problem; one meets a similar problem in other, older
optical systems. For example, if you pull the field lens out of an
eyepiece of a microscope or telescope you will observe the same
effect. The field lens just collects all those rays which are
BEYOND falling directly onto the eyelens and bends them inward so they do
reach the eyelens.
In fact, this leads me to something I wanted to mention; if you place the
right Fresnel lens on the scanner glass it can do the job of the field lens
and repair the "only a bright spot" effect. This also works when there is
a ground glass present, though much less effectively because the ground
glass scatters the light into all directions and the field lens cannot fix
that. Before you all run out looking for the right Fresnel lens I should
remind you that it will cause rings to appear in your scans, though you can
certainly reduce the intensity of those rings by placing the Fresnel lens
somewhat away from the image plane (this is also an old problem; some
eyepieces have the field lens at the focus of the telescope or microscope
objective and you get a nicely magnified view of each dust particle sitting
on the lens.) I have not worked this through but I believe (please do not
blame me for a claim this is fact; it is just something to think about and
maybe one of you will work out the right answer before I do) you need a
more complicated Fresnel lens consisting of a conventional one behaving
like a lens with a spherical surface plus another one which functions as a
cylinder lens, something which is possible to make but which I have never
seen. This is the result of the direction in which the scanner lens looks
as it moves across the film plane. You could use a plain vanilla Fresnel
lens if the scan were made by rotating the scanner head in place to view
the image rather than moving it across the film plane while it always
points looking out in a direction perpendicular to the glass, that is, the
way a usual scanner works. (Again, nothing new here, this is the panorama
camera looking at the output of the Fresnel lens rather than the real world
directly! There is little to recommend this system except for educational
purposes.)
Bob
you wrote:
>For those who may question Bob's assertion that one would only get a small
>image circle if a lens is placed above a scanner, I admit I did this
>experiment a while ago. It is as he and others claim.
>
>I did not keep a copy of the scan, I figured it was just trash.
>
> I used a 8 X 10 camera and just tilted the scanner against the open back of
>the camera with the lens open. Lamp on, nothing special to trick the
>scanner.
>
>Mostly black except for a small circle shaped image in the middle of what
>the camera was pointed towards.
>
>Try it yourself. Maybe the same thing can be found with your 35mm camera.
>Place it on a flatbed scanner looking up. Let us know.
From camera makers mailing list:
Date: Wed, 14 May 2003
To: cameramakers@rosebud.opusis.com
From: Robert Mueller r.mueller@fz-juelich.de
Subject: [Cameramakers] Why you cannot just put a scanner in place of a large format sheet film holder
Below you will find an Internet address for the site of Lester Hawksby, who
very kindly wrote to me and offered to host the drawing I expressed the
desire to make available about a common misunderstanding in trying to put a
flatbed scanner behind a large format camera to make full size electronic
images. He also did a nice job of cleaning up my hurried effort and he put
it into pdf format, more appropriate for you than my original Word document.
http://www.bits.bris.ac.uk/lester_hawksby/photo/scannerdiagram.pdf
I hope this helps to explain what is going wrong and maybe encourage some
experiments which avoid the problem and thus end up making fine images (and
telling us how to follow in your footsteps!)
Thank you Lester!!!
Bob
From camera makers mailing list:
Date: Wed, 14 May 2003
From: Philip willarney pwillarney@yahoo.com
To: cameramakers@rosebud.opusis.com
Subject: [Cameramakers] scanner kludge camera
Re: the question about building a scanner from a
camera (oops, other way 'round) I saw this a while
ago, and put it on my list of
I-could-do-that-someday-I-will projects..
http://www.sentex.net/~mwandel/tech/scanner.html
Date: Thu, 15 May 2003
From: huw huw@huws.homelinux.net
To: cameramakers@rosebud.opusis.com
Subject: [Cameramakers] Scanning back camera from a scanner
Hi,
For details of some scanners 'commands' try the Linux lot, they have
either reverse engineered the interface or (in the case of Primax at
least) got the 'real' interface specs.. All source code and probably an
explanation of how it works are available on the web.
Huw
From minolta mailing list:
Date: Fri, 13 Jun 2003
From: "Peter Blaise Monahon peterblaise@yahoo.com" peterblaise@yahoo.com
Subject: Re: DiMage Scan Dual 3 Scanner - low math version of dynamic range
> Date: Mon, 09 Jun 2003
> From: --_-----_-------@-----.----
> Subject: DiMage Scan Dual 3 Scanner
>
> I'm interested about buying this
> scanner, but I have a few
> questions. According to the specs,
> it has a dynamic range of 4.8
> (max). Is this accurate...
Yes, 4.8 dynamic range.
> ...or do they
> use some software to enhance
> color the same way interpolation is
> used to enhances resolution?...
No. It is a straight scan, no interpolation.
> ...What is it's true dynamic range?...
4.8 dynamic range as claimed.
> ...It sounds too good to be true ...
Yes. Ain't Minolta grand?
> ...The scanner also has a 16 bit
> conversion feature. Does this mean
> you can make a 48 bit scan?...
Yes. And it's not a conversion feature, it's a
direct 16bit scan.
> ...Any comments are welcome since
> I'm new to this subject area.
>
> Thanks,
>
> ----
Hello fellow Minolta User's Groups members,
After at least a 4 day break to let the steam out of this
thread from the main Minolta list, I hereby revisit it.
Please let me know if I stay on target: our experience of our
gear and our photography, and helping each other with our
experience of our gear and our photography. This thread has
also been called [RE: DiMage Scan Dual 3 Scanner - low math
version of dynamic range], so I will copy this to both threads
for our archives on the main list.
--------------------------------------------------------------
[1] REVIEWER'S EXPERIENCE:
The MDSDIII Minolta Dimage Scan Dual III has been getting
rave reviews after hands-on experience in the photography
magazines - reviews from people like you and I, only they get
paid! I've already shared the Popular Photography magazine
review
http://groups.yahoo.com/group/minoltascandual/files/minolta-dsdiii-pop-ptoto-april-2003-900s.pdf
and now there's a Digital Photo magazine mini-review (the
magazine itself also includes test pictures and full
specifications on CD)
http://groups.yahoo.com/group/minoltascandual/files/minolta-dsdiii-in-digitalphoto-2003-03-900s.pdf
and a Photographic magazine review
http://groups.yahoo.com/group/minoltascandual/files/minolta-dsdiii-photographic-review-may-2003-900s.pdf
along with various other reviews world wide. The most salient
quote I can share is this:
"...the [Minolta Dimage Scan Dual III] scans are
certainly as good as those from our much more costly
4000-dpi scanner up to the III's rated 2820 dpi,
with excellent shadow and highlight detail,
sharpness and color fidelity..."
[Complete text below * ] No one ever said anything like
this about the MDSDII Minolta Dimage Scan Dual II. The
significant differences between the II and the III are bit
depth and dynamic range, resolution remains the same:
MDSDII 2,820ppi 12bit 3.6 dynamic range
MDSDIII 2,820ppi 16bit 4.8 dynamic range
--------------------------------------------------------------
[2] USER EXPERIENCE:
There is a Minolta User's Group dedicated to the Minolta
Dimage Scan Dual scanners, and there, you can view and share
and discuss hands-on experience and see some FANTASTIC scans
from the real experience of the users - approximately 128
members. They tend to agree about the maximum image quality
differences capable from the III over the II. Please visit
there, join, and see for yourself.
http://groups.yahoo.com/group/minoltascandual/
Again, the significant differences between the II and the III
are bit depth and dynamic range, resolution remains the same:
MDSDII 2,820ppi 12bit 3.6 dynamic range
MDSDIII 2,820ppi 16bit 4.8 dynamic range
--------------------------------------------------------------
[3] MY EXPERIENCE:
Comparing my own scans on the older model II to those of
my fellow scanners using the newer model III, I fell "shock
and awe" - their scans are that much better than mine. For
another user's hints on maximizing the older model II, you can
check out http://www.willegal.net/photo/dual-ii/dualii.htm
Again, the significant differences between the II and the III
are bit depth and dynamic range, resolution remains the same:
MDSDII 2,820ppi 12bit 3.6 dynamic range
MDSDIII 2,820ppi 16bit 4.8 dynamic range
--------------------------------------------------------------
I hope this puts into perspective some of the previous
speculative theory offered earlier in this thread, postulating
that increased bit depth and dynamic range could not possibly
contribute to better image quality, and that only an increase
in resolution could improve image quality - speculation not
based on experience of our gear or our photography.
When comparing two products, the goal is to get "all
others things being equal", and in this rare case, we have a
clean example where all other things really are equal.
And, in this case, resolution has not changed, only bit
depth and dynamic range have increased, and, not surprising to
me, the image quality has improved - and this is based on
hands-on experience of our gear and our photography.
And not just a little improvement.
Image quality has improved to be the equal of scanners
that cost thousands of dollars more, when compared at 2,820
ppi.
Wow.
Got Minolta?
Peter Blaise Monahon - peterblaise@yahoo.com -
http://www.geocities.com/minoltphotostudentsandteachers/
The Mind of Minolta is a terrible thing to waste!
--------------------------------------------------------------
PS - * The complete text of the Photographic magazine review
of the Minolta Dimage Scan Dual III:
User Report
Minolta Dimage Scan Dual III
Repro-quality scans of your slides and negs, for
about $300!
by the Editors of Photographic Magazine, page 52,
May 2003 http://www.photographic.com/
If you've been thinking about "going digital,"
but your budget doesn't allow for a "serious"
digital camera or one of those high end film
scanners, this is the item you need.
Minolta's DiMAGE Scan Dual III is a $300 film
scanner that turns your 35mm (and, with optional
adapter, Advanced Photo System) negatives and slides
into repro-quality scans for use in your computer.
You can "go digital" with all your existing film
images, and continue using your familiar film
camera. As its name suggests, the "III" is the
third-generation version of the "economy' member of
the company's deservedly well respected film-scanner
family. But "economy" doesn't mean lacking in
performance or features. The "III" has a maximum
resolution of 2820 dpi, which is high enough to
publish a scanned 35mm image bigger than full-page
in this magazine at 300 dpi (and more than ample to
produce great 13x19-inch inkjet prints). It's got
an amazing dynamic range of 4.8, and while published
dynamic range figures for scanners tend to be a tad
controversial, in our tests the III's scans did give
us all the shadow and highlight detail our original
negatives and slides contained. The color depth of
16 bits produces clean, noise-free scans with good
color and detail throughout, and 2X, 4X and 8X
multi-sampling further improves image quality. The
III also offers automatic dust removal (via
Minolta's own Auto Dust Brush rather than the
Applied Science Fiction Digital ICE feature that
accompanies higher-end Minolta scanners), instant
correction of off-color and faded originals (via
Minolta's Pixel Polish software), quick scan times
via USB 2.0 (it also works with USB 1.1 if your
Windows or Mac computer doesn't have USB 2.0), and
very easy operation.
In Use
You can use the DiMAGE Scan Dual III with both
Windows PCs (98 and up, Pentium 166MHz or faster)
and Power Macintosh computers (OS 8.6 through OS X,
G3 or later processor). The manual says you need at
least 64MB of RAM, but as with all digital-imaging
products, the faster the computer and the more RAM
you have, the better.
Once the software is installed (a simple matter
of clicking through some on-screen instructions),
you just connect the scanner to the computer with
the supplied USB cable and to an AC outlet via the
supplied DC adapter, switch it on, and open the
scanning program. Minolta provides two choices
here: the Easy Scan Utility for those new to
scanning, and the DiMAGE Scan Dual III Utility for
"power users" who like full control over everything
(this advanced utility can operate as a stand-alone
program or as a plug-in for your image-editing
program). Once the desired utility is up and
running, load the film, and you're ready to go. The
scanner comes with holders for six-frame 35mm
negative strips and four mounted slides (the
film-strip holder is one of the easiest we've
encountered in a film scanner), and you can add an
adapter for APS films for around $150.
[Image]
[Image] The DiMAGE Scan Dual III does good work:
This 2820-dpi scan from a 35mm negative of a night
scene shows lots of detail throughout and easily
holds up published at full-page size. And the colors
are excellent.
[Image]
[Image] The Scan Dual III handles slides as
easily and as well as negatives (and actually scans
them faster, since it doesn't have to convert the
image from negative to positive). Dynamic range,
colors and detail are excellent.
[Image]
[Image] Here's another tough image for a
scanner. The DiMAGE Scan Dual III handled it
extremely well.
With the Easy Scan Utility, click the
appropriate on-screen film-type button (color
negative, B&W; negative, color positive, B&W;
positive), then click the Next button. The scanner
will proceed to make index scans of the loaded
images, which will appear on-screen when the
scanner's done. Click the frame you wish to scan
(only one image at a time can be selected in Easy
Scan), and click the on-screen buttons to do any of
the following: Pixel Polish (color film only -
automatically corrects images), Auto Dust
(automatically reduces the effects of dust), rotate
counterclockwise or clockwise, rotate all frames
180', and adjust image (which brings up another
screen, on which you can adjust brightness, contrast
and saturation via sliders). Then, select the final
use of the image (view on screen, use for printing,
use as material for processing), and click the Scan
button. A save-as screen will appear, via which you
name the file, choose a format (BMP, JPEG, TIFF or
PICT), and tell the program where to store the
scan. Then click the Scan button, and the scanner
will make some strange noises (all scanners make
strange noises!) and put your scanned image where
you told it to put it.
With the DiMAGE Scan Dual III Utility, you can
do simple scans as with the Easy Scan Utility, but
you can also do a lot more, including flip the image
horizontally or vertically (or both), magnify the
prescan image, crop the image, adjust the
resolution, and exercise precise control over
exposure, color and contrast via curves and
histograms. You can choose the color space (sRGB,
the popular Adobe RGB and lots of others), use ICC
profiles, save and apply settings to other images,
batch-process all six film-strip images or all four
slides, and more. It's all easy to do, via
on-screen buttons and tabs.
Bottom Line
The DiMAGE Scan Dual III is an ironic case where
the product's low price actually kind of works
against it: We initially had our doubts about it
because of its low price (how good could a $300 film
scanner be, anyway?). Well, we were very pleasantly
surprised -- the scans are certainly as good as
those from our much more costly 4000-dpi scanner up
to the III's rated 2820 dpi, with excellent shadow
and highlight detail, sharpness and color fidelity.
The III is easy to use, gives you plenty of control
when you want it (the auto scans are very good),
performs surprisingly quickly, and turns out great
scans. What more can one ask from any scanner, much
less a $300 one?
Dimensions are 5.7x3.9x12.6 inches (it takes up
a lot less room than a flatbed scanner!) and 52.9
ounces. For more info, go to
http://www.minoltausa.com/ or use Minolta's PHOTOFAX
at 1-800-528-4767.
From: Mxsmanic mxsmanic@hotmail.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: $2 "coke bottle lens" on scanners vs. Zeiss glass & film
Date: Fri, 13 Jun 2003
Bob Monaghan writes:
> Most seem to be about $2 mfgers cost items (granted,
> the scanners probably cost $50 to make for a $300
> retail scanner ($150 wholesale, $75 importer..), little
> more than an uncoated cylindrical lens.
Judging by results, I'd say that Nikon, at least, uses some really good
lenses in their scanners. The 8000ED uses a 14-element, 6-group lens
with 6 ED elements, which implies that some significant work has gone
into its design and manufacture. Other Nikon scanners are comparable.
Of course, not all vendors necessarily do this. And a cheap lens would
certainly have a very negative effect on scanner performance, just as a
cheap lens has a very negative effect on original photography.
> Folks spend kilobucks on zeiss lenses in MF, then
> scan the film with a $2 piece of bottle glass ;-)
Scanner lenses may have somewhat more limited requirements, though,
since they only need to focus a thin line, and not a whole
two-dimensional field (?).
--
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.
From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: $2 "coke bottle lens" on scanners vs. Zeiss glass & film
Date: 13 Jun 2003
Mxsmanic mxsmanic@hotmail.com wrote
> Scanner lenses may have somewhat more limited requirements, though,
> since they only need to focus a thin line, and not a whole
> two-dimensional field (?).
CCD scanner lenses are conventional rotationally symmetric lenses, and
they must therefore cover a full circle whose diameter is equal to the
small dimension of the scanned object. The advantage is that this
dimension is only about 60% of the full diagonal of the film frame.
This is where the more limited requirement comes about, and it is
significant.
Brian
www.caldwellphotographic.com
From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: $2 "coke bottle lens" on scanners vs. Zeiss glass & film
Date: 14 Jun 2003
Stacey fotocord@yahoo.com wrote
> Thanx for the info, I had been lead to believe that when scanning film the
> scanner only had to focus on each pixel at a time.
Hi Stacey:
What I said is true for scanners using a linear (or tri-linear) CCD.
Drum scanners do only digitize a single point at a time, so
microsope-type objectives can be used. This is the main reason why
drum scanners can achieve such high resolution. I think that some
commercial drum scanners go as high as 11,000 dpi (216 line pairs per
millimeter).
Brian
www.caldwellphotographic.com
Date: Wed, 18 Jun 2003
To: contax@photo.cis.to
From: Joe Doehler contax@doehler.us
Subject: Re:[Contax] Scan Resolution
you wrote:
>The Epson Perfection 3200 had the highest Dmax of all tested scanners
>including Film-Scanners! Resolution was somewhere around 2400dpi if I
>remeber correctly.
>...
This site compares scans from the Epson 3200 and the Nikon 8000ED:
http://www.digit-life.com/articles2/epson3200/index.html
You can judge for yourself.
Joe.
From: "Austin Franklin" darkroom@ix.netcom.com
To: contax@photo.cis.to
Subject: RE: Antwort: [Contax] Scan Resolution
Date: Thu, 19 Jun 2003
Michael,
> The Minolta scanner actually has a lot going for it.
> A Dynamic range of 4.8 is extremely good,better than my drum scanner in
> fact !
I HIGHLY doubt that is true. A dynamic range of 4.8 simply means that the
scanner uses a 16 bit A/D, NOT that the scanner can actual get 16 bits of
clean data. If it could actually get in the high 3's, I'd be highly
surprised. This is merely abuse of specs by marketing.
> The 3 line CCD sensor is as good as you can get with one line each of red
> green and blue for purer colours.
That does not give "purer colors". ALL scanners not a days use a tri-linear
CCD (three lines).
> Minolta have been at the forefront of things electronic in the photo arena
> for a long-time and know what they are doing.
I question that... The other Minolta scanners though OK, weren't anything
spectacular, and certainly no better (if not worse) than other manufacturers
comparable models. What Minolta has done is pushed the "spec" envelope, not
necessarily made better scanners, but scanners that they CLAIM are better by
abusing the definition of commonly used specs and terms.
Austin
From contax mailing list:
From: "Austin Franklin" darkroom@ix.netcom.com
Subject: RE: Antwort: [Contax] Scan Resolution
Date: Thu, 19 Jun 2003
Hi Pat,
Grain aliasing is primarily caused by the type/design of light source and
lense. Obviously, the film has a lot to do with it (as does the actual
scanner resolution), but only from the standpoint that certain films react
with the light/lense combination worse than others. Some scanners just
don't have this effect nearly as much as others, with the same film.
Regards,
Austin
From contax mailing list:
From: "Austin Franklin" darkroom@ix.netcom.com
Subject: RE: Antwort: [Contax] Scan Resolution
Date: Thu, 19 Jun 2003
Hi John,
> Isn't scanning at a resolution greater than the grain a way of avoiding
> grain aliasing? In fact by scanning at a higher resolution and bit-depth
> than is required and then reducing resolution/bit-depth after editing, and
> appropriate for the end use, I get better quaility images.
Scanners have a native optical resolution that they scan at, no matter what
you select it to scan at. This means that if your scanner is a 4000SPI
scanner, and you choose 2700...it'll still scan at 4000, and downsample the
image data (therefore degrading it) to 2700. My recommendation is simply to
scan at the optical resolution of the scanner, and when printing to inkjet
printers, simply resize the image WITHOUT resampling (uncheck the resample
box in PS) and send what ever PPI it results to directly to the printer
driver.
Austin
From: "David J. Littleboy" davidjl@gol.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Scanner comparison-Epson 2450 vs Nikon 8000
Date: Sat, 10 May 2003
"Rick Rieger" rrieger@wideopenwest.com wrote:
> Now for the interesting part...I could tell very little difference in detail
> resolution between the two scans, nor was there much difference in apparent
> dynamic range. The image was of a lighthouse in the sun with bright
> highlights and deep shadows. Actually, the Epson scan was much sharper at
> the edges of the frame. There may have been a film flatness issue with
> the Nikon.
Your friend wasn't using the glass carrier. The Nikon's DOF is about +/- 10
focusing units (in particular, sharpness is noticeably worse 20 units off
from the measured focus position), and I've not been able to consistently
hold focus across the whole frame without it. Some people claim to get away
without the glass carrier, but I find it hard to believe.
> I don't know if the Nikon scanned the image any faster, because I wasn't
> present for this scan.
>
> Admittedly this is just one comparison test, and it would not be prudent
> to draw conclusions from it.
Exactly {g}.
> What I have decided is that for now, the Epson
> will do fine for me on my MF scans.
IMHO, from 6x7 and larger, the 3200 should keep a lot of people happy. I
shoot 645, and found the 2450 wasn't getting enough detail from my slides to
make 645 competitive with the dSLRs.
> I have done some 4x5 scans on the 3200 with really excellent results.
> Right now, I'm a happy camper.
David J. Littleboy
Tokyo, Japan
From: "Rick Rieger" rrieger@wideopenwest.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Scanner comparison-Epson 2450 vs Nikon 8000
Date: Fri, 9 May 2003
I recently compared a scan of a 6x7 E100VS image from my Epson 3200 against
a scan of the same chrome from a friend's Nikon 8000. The Epson scan was
done at 3200dpi, and the Nikon scan at 4000dpi. Opening both in Photoshop
revealed that the Nikon scan was much closer in color balance to the
original than the Epson scan. When the Epson scan was made however, I did
not have any color management applied to my system. Now I do with the
Monaco Optix system. The Epson scans are now better matched against the
originals. In any case, I was able via PS to adjust the color, brightness,
and contrast of the Epson scan to be very close to the unaltered Nikon scan.
I am using Silverfast 6i scanning software on the 3200.
Now for the interesting part...I could tell very little difference in detail
resolution between the two scans, nor was there much difference in apparent
dynamic range. The image was of a lighthouse in the sun with bright
highlights and deep shadows. Actually, the Epson scan was much sharper at
the edges of the frame. There may have been a film flatness issue with the
Nikon.
I don't know if the Nikon scanned the image any faster, because I wasn't
present for this scan.
Admittedly this is just one comparison test, and it would not be prudent to
draw conclusions from it. What I have decided is that for now, the Epson
will do fine for me on my MF scans. As time goes on, I will try to do more
comparisons and watch the marketplace. The Nikon 8000 has recently declined
in price by several hundred dollars.
I have done some 4x5 scans on the 3200 with really excellent results. Right
now, I'm a happy camper.
FWIW and YMMV.
Rick Rieger
josburke@bellsouth.net wrote
> Before you recommend the Epson 3200 please read through. I have an Epson
> 2450 that really does a decent job at scanning MF images. My question is HOW
> MUCH BETTER IS THE NIKON 8000??? HOW MUCH FASTER? I chose the Epson 2450 at
> the time because it is/was reasonably priced AND can scan a 4x5 as well
> which I also use. I'm really not looking to spend big bucks right now but
> this is just a matter of curiosity.
From: "David J. Littleboy" davidjl@gol.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Scanner comparison-Epson 2450 vs Nikon 8000
Date: Sat, 10 May 2003
josburke@bellsouth.net wrote
> Before you recommend the Epson 3200 please read through. I have an Epson
> 2450 that really does a decent job at scanning MF images. My question is HOW
> MUCH BETTER IS THE NIKON 8000???
Worlds. Especially for slides. If the user gets the whole frame in focus,
which (in my experiences) often requires using the glass carrier, which ups
the cost another US$250.
> HOW MUCH FASTER?
Not much. On the 8000, a 645 scan takes from 5 minutes (the way I scan all
my slides) to very long (for 16x scans, which are, in my experience, not
required).
> I chose the Epson 2450 at
> the time because it is/was reasonably priced AND can scan a 4x5 as well
> which I also use. I'm really not looking to spend big bucks right now but
> this is just a matter of curiosity.
The 2450 has minor trouble with contrasty slides (a flare-like effect from
bright areas) and is more than a tad soft. But it's good enough for at least
5x enlargements for lovely 11x14s from 6x7 negatives. Beyond that, it's a
matter of how fussy you are. The 3200 provides slightly improved resolution
and faster scans. Even if you're thinking about getting the 8000, you should
get the 3200 first and (a) learn how to scan, (b) determine if you can stand
the incredible waste of time scanning is, and (c) assure yourself that you
really need to spend all that money.
David J. Littleboy
Tokyo, Japan
From minolta manual mailing list:
Date: Tue, 24 Jun 2003 10:22:28 EDT
From: BillB800si@aol.com
Subject: Re: Example: flatbed vs film scanner
bxfitz@dodo.com.au writes:
> Are you scanning prints or negatives on your flatbed scanner?
Couple of months ago I took a two day class with George Lepp
http://www.leppphoto.com/
and the only flat bed scanner he said he uses is Epson's new
3200.
He also mentioned that he had run a series of tests using the Nikon and
Minolta scanners and Nikon won out.
A few other hits he gave:
1- Uses Epson printers with Pigment Inks not Dyes. Mentioned that the Epson
1270 and 1280 prints sometimes degrade because of Ozone (not UV light).
2- Usually uses 1440 DPI not 2880 DPI.
3- Velvia not best for wildlife shots, he uses Provia.
3- He always avoids using 1/15 second shutter speed. (??)
4- Most of the time he uses "Spot Metering".
5- Uses Nikon 5T lens- don't lose light.
6-Kenko Extensions maintains AF feature of camera. Uses 12mm for most of his
Macro shots.
Finally, scan at the highest resolution that your scanner allows and highly
recommends ICE feature, if it has it.
Just a few thoughts I picked up at his seminar.
Bill B. (USA)
From minolta manual mailing list:
Date: Sun, 22 Jun 2003
From: "Mary Hurst" mhurst@pineland.net
Subject: Re: Re: Scanners and Printers
My flatbed scanner scans negatives . . is that what you are talking about?
----- Original Message -----
From: Paul Brecht
To: Minolta@yahoogroups.com
Sent: Sunday, June 22, 2003
Subject: [Minolta] Re: Scanners and Printers
Mary,
For quality control, I'd suggest a film scanner. I use the Minolta
Dimage Scan Elite II. (about $700 @ B&H;) It works wonderful, but I do
notice that I top out at 8x10.
Pretty soon (any time now) Minolta will release a higher resolution
model with more options & features for a little more ($800-900). This
seems to be the ideal model. Minolta scanners are better at color
rendition & have a wider dynamic range then the competition...
If that's too much for you, Minolta has the Scan Dual III that costs
around $300-400...
To be honest, as scanner isn't something you'll want to do window
shopping by foot on. The best bet is to do research on-line & buy
from an on-line retailer like B&H; Photo & Video...
http://www03.bhphotovideo.com/
The reason I recommend a film scanner is that it puts all the control
back into your hands. You don't have to trust your images to an
underpaid enployee (or as Peter stated "STORE CLERK") who really
doesn't care about imaging, quality or even how to use the print
machine the best ways possible...
The problem with flatbed scanners is that you will be starting with
an image that was altered by the "store clerk" & usually given a 1/2
decent print. Another issue is that the print paper is only capable
of holding hundreds of colors while a Minolta film scanner will
capture millions of colors. Well, film holds thousands of colors &
the film scanner is capable of capturing every color available in the
film...
Most scanners (& digital cameras) come with Adobe Elements. This is a
stripped down consumer version of Adobe Photoshop. Elements is more
than ample for correcting color issues, cropping, etc...
I made my decision to buy a mid-range scanner (Elite II) because I
went & got some slide film scanned at a pro shop & found out that I
could easily get my money back in 20 rolls worth of high-res CD's.
That was 300 rolls ago...
I'd suggest getting the new 5400 scanner as your best case scenario.
The Elite II as 2nd best & Scan Dual III as 3rd. Best bet - take tour
time & don't rush. Look into it & make the best informed decision you
can. Don't get rushed by means of GAS (gear acquisition syndrome) -
the more well educated your purchase, the better investment you'll be
making...
This principle should be applied with all you gear purchases.
Consider them an investment. Buying the product that's within reach
now isn't always the wise investment. Usually it's thinking about how
you'll never purchase another _________ (fill in the blenk) ever
again & you want to make sure that you have the right one (or you'll
end up paying for another one because you bought wrong)
Sometimes it's cheaper to buy more expensive...
Paul
--- In Minolta@yahoogroups.com, "Mary Hurst" mhurst@p... wrote:
> Several of you have mentioned that you scan your own photos in,
edit them, and print. What scanners and printers do you recommend?
Considering that I have to drive over 20 miles to get to the nearest
store of any kind, looking into this in the future might be worth my
while.
>
> Mary
From manual minolta mailing list:
Date: Sun, 22 Jun 2003
From: "Paul Brecht" pariht@pacbell.net
Subject: Re: Scanners and Printers
Here's some related info:
http://www.scantips.com/basics13.html
http://www.cix.co.uk/~tsphoto/tech/filmscan/
http://www.steves-digicams.com/scanners.html
http://www.imaging-resource.com/SCAN1.HTM
Paul
From minolta manual mailing list:
Date: Sun, 22 Jun 2003
From: "Chuck Cole" cncole@earthlink.net
Subject: RE: Re: Scanners and Printers
Hi Mary,
I chose a factory refurbished HP S20 USB model off eBay for $160. HP specs
are trustworthy physical measurement numbers while others may not be. HP is
a great product line, and it comes with good software tools. This scanner
has a modest resolution of 2400dpi (dots per inch... That's 17 megabyte
files at max resolution). This unit has a real optical dynamic range of 3.2
which, as a true physical measurement, is more than any film can actually do
(this is documented).
Paul and I are on different tracks of scanner measurements somehow. Not
sure who may be off or where, and my electro-optical physics is pretty good.
Most recent scanner ads quote a "numerical dynamic range" which is BS and
not the actual optical capacity of the ccd system. This is only a
translation of the digitizing bits... as if using really tiny markings on a
ruler had some effect upon measuring the "jiggling bowl of jello" we call
film. There are practical reasons why these numbers may apply to the
scanner somehow, but it's marketing hype to interpret them as "dynamic
range" as if that number applied to some useful film range. All of the film
numbers are somewhat uncertain and slightly inconsistent, like jello
wiggling a little. I can't agree with Paul: any claim of dynamic range over
3.2 is unlikely to be a real or meaningful physical number about the ccd's
optical range of film scanning, so it's hardly a "market lead" in real ccd
capability. I'd love to be shown to be naive or mistaken on this, or to
discover that I'm overlooking something. All claims for a dynamic range
like 4.8 I can find are pure marketing hype and only a translation of the 16
bit digitizing number: none trace to optics and/or film properties as the HP
numbers do. Furthermore, those extra bits are discarded before making the
output file to save file size. If the numbers were "true" and applicable, a
4000dpi film scanner with "dynamic range" of 4.2 would always produce scans
of 1.15 gigabyte size per photo: anything less would be losing info, if that
info were truly there. They are not claims for having factors of ten more
gray scale range, etc, as the numbers should represent if they were in fact
a photographic dynamic range.
I decided that 2400 dots per inch (roughly 8 megapixels if this were a DSLR
spec) is plenty for about 90% of what I do and I thought I could go to a
nearby pro-lab when I needed any more. I later found out that the best
local pro-lab doesn't offer any higher than 2400 on their pricing sheets,
and they want high $$ for just that. They do not offer any higher
resolution, but were willing to try to locate an outside service for me.
This local pro-lab is a big one that's online also, so it's not easy to
locate anything better in any big city. I think finding better custom
digitizing work from another pro-lab would be hard, but I haven't checked
yet.
One of the low-end Minolta scanners has similar resolution to my HP S20 and
can be bought for under $100 as a factory refurb off eBay. However, I much
prefer the way the HP S20 handles film and how the HP hardware and software
work with my HP computer. Minolta is not a computer manufacturer, nor do
they have the extent of user support options HP has. I've been very
disappointed at the lack of tech support most big camera shops have today.
I was unable to get basic Minolta Maxxum lens compatibility questions
answered by 4-5 big online shops I asked, and one actually stated that they
have no technical support people at all. I would not expect any useful info
or support for a scanner from a camera shop: buy wisely and rely upon
manufacturer's support if any is ever needed. My HP S20 refurb came in a
sealed box with HP warranty and support.
My printer is an Epson 1520 which can do 17x22 paper sizes in photo quality
(1440x720dpi). It cost $495 new and does any paper size up to 17x22. Good
ones are common on eBay now. I haven't fully evaluated this printer for
photography, though I've had it quite a while. The HP and Epson printers
which can do the 13x19 paper size are newer and better, so I might go for
the newer versions of the HP 1220CSE. HP models are most popular, so are
best for getting a wide range of photographic inks and papers beyond the
factory stuff and stuff at OfficeMax, etc.
YMMV..
Chuck
From Minolta Mailing List:
Date: Sun, 22 Jun 2003
From: "Paul Brecht" pariht@pacbell.net
Subject: Re: Scanners and Printers
I've heard these arguements before...
While there may be some validity issues to the actual numbers
(specifics), it has been proven though that the mythical Minolta 4.8
numbers really do have more color than it's Canon & Nikon
counterparts... (which claim 3.2 (N) & 3.6 (C))
I'll tell you what the limiting factor is: colorspace...
When I scan, I scan in Minolta's Wide-Range RGB gamut & the colors
are phenomenal. I notice when I open it up in Adobe, I lose a lot of
color because Adobe has a limited colorspace (Adobe RGB) When I save
to jpeg, I see a greater loss & when you view the image through any
Microsoft product, you still lose more color. Microsoft is bent on
the generic sRGB, hence you lose lots of color...
When I look at my image on the Dimage Scan software, the colors are
unworldly when compared to what anyone will ever see on the web...
With Minolta's scanners, you have the option for scanning at 16 bit.
That's 16 bit for each color which translates to 48 bit color
sampling in RGB RAW or TIFF format...
I currently am scanning at 8 bit because in certain situations, I
don't want too much color & find myself spending a lot of time
desaturating images that don't need it when I scan at 8 bit...
Paul
From: Rafe B. rafeb@adelphia.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Pentax 6x7 - why?
Date: Fri, 27 Jun 2003
Jeff tokom@sympatico.ca wrote:
>At this point I can't see myself spending $3000 for one of the dedicated MF film
>scanners.
Check out used MF film scanners on eBay.
There's a LeafScan 45 selling for $899 ("Buy it now")
with current bidding at $699.99 (no bids yet, 4 days to
go.)
I was seriously considering buying one of these
a couple years back. It'll give you a very honest
2500 dpi, with superb optics. 5000 dpi on 35 mm.
There's also a umax Powerlook 3000 at $200
with 4 days left (3048 dpi.) That's rather an
incredible price -- this was a $6000 scanner
just four or five years ago.
Other models to look for: Agfa T2500, SprintScan
120 (LS-120.) None should cost more than $1K
nowadays, pre-owned.
rafe b.
http://www.terrapinphoto.com
From: Rafe B. rafeb@adelphia.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: only 6x8" prints from 645? Re: Are used MF...
Date: Sun, 08 Jun 2003
rmonagha@engr.smu.edu (Bob Monaghan) wrote:
>But again, the data is there on the film, with the right scanning
>technique. And it can be recovered today, though not with low end scanning
>gear used by most prosumers today. In the future, I assume improved film
>scanning technology will make it (painfully) clear how much more quality
>is recorded in today's films. You can get that quality today, though the
>cost is moderate ($15-50+/drum scan), the pro quality is there and obvious
I used to believe that, but not any more.
See http://homepage.mac.com/anton/NikonTango/
In short: the Nikon 8000 (CCD) doesn't fare badly at all
compared to one of these high-end drum scanners.
With good optics and 4000 dpi, you're capturing most of
what's available on the film.
Minolta's newest 35 mm film scanner, due out in a matter
of days, resolves 5400 dpi. It will be interesting to see
the output from that machine. OTOH, I have seen the output
from a Leafscan (5000 dpi on 35 mm.)
Possibly those shooting with the finest German lenses,
under ideal conditions, may find more detail in their film
scans than I have in mine.
rafe b.
http://www.terrapinphoto.com
From: Bob Monaghan [rmonagha@engr.smu.edu]
Sent: Sat 6/7/2003
To: Monaghan, Robert
Subject: Re: Are used medium format camera prices dropping?
I'm forced to agree (again) with John S.- the big problems I keep reading
about on many scanners is they have very little depth of acceptable focus
when scanning film, and if you don't use the right accessory carrier
properly you can easily get outside the acceptable DOFocus limits (which
can be well under 1 mm (as in 0.2mm) on some scanners. Much of the poor
quality in scanning seems to be a side effect of out of focus image losses
due to problems with maintaining flat film in these narrow DOFocus zones.
This problem is also present in enlargers, but enlargers have lenses that
you can easily stop down and generally do in making sharper prints. You
can't do that on the integrated optics on a scanner. That means your
enlarger has a much larger DOFocus when stopped down. Moreover, flat film
carriers, and especially glass plate carriers, make it possible to get
rather better results from enlarged film than by scanning it.
my $.02 ;-) bobm
From: Mxsmanic mxsmanic@hotmail.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Are used medium format camera prices dropping?
Date: Fri, 06 Jun 2003
Rafe B. writes:
> Resolution and detail are not at all the same thing ...
Actually they are.
> You cannot resolve details (using a scanner)
> that don't exist on the object being scanned (ie., the film.)
Correct.
> The resolving power of film is finite, as is the
> resolving power of any digital capture device. Both
> are limited by optics, of course -- in that regard
> they are identical.
Correct.
> I urge you to do the test yourself, if you own or have
> access to a film scanner.
I have three scanners, and I scan film every day.
> Compare high-quality scans of your very best 6x6
> transparencies against images from the EOS 1Ds.
I have. My very best transparencies--and indeed, even my mediocre
transparencies--leave the 1Ds and every other digicam so far in the dust
that it makes me smile each time I compare. Even 35mm transparencies
look better.
> Believe me -- I was quite surprised at the results.
I wasn't. The only thing that surprises me is that so few other people
see this. I suspect, though, that most people have no clue as to how to
properly scan film.
--
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.
From: fcarello@tiscalinet.it (Fernando)
Newsgroups: rec.photo.equipment.large-format,comp.periphs.scanners
Subject: Re: Relation of bit depth to dynamic range
Date: 15 Jul 2003
Kennedy McEwen rkm@nospam.demon.co.uk wrote
> years and has, in the main, been propagated in this thread. Also note,
> as few here appear to have, that noise is not the minimum signal, but
> the minimum discernible difference between signals according to well
> defined mathematical and physical criteria - and the CCD just happens to
> have some very complex noise characteristics, being a combination of
> constant, linear and exceedingly non-linear terms. CCD signal may well
> be linear, noise certainly is not, and that can be very significant when
> comparing slide and negative scanner performance.
Very interesting as always.
Could you clarify a bit this aspect (CCD noise characteristics and
their relationships with actual neg and slide scans)?
Also, I'd like to point out (not to you, but in general) that is not
true that slides have a 3.0-3.3 max density; just looking at Velvia 50
charts as provided by Fuji, it's evident that this film can sport as
high as 3.7D at low exposures, and the new Velvia 100 seems to get
even densier (near 4.0D).
So it's very difficult even for a good scanner to catch all the
dynamic range of a contrasty Velvia slide, and Velvias are among the
most used slide films... and this, without even taking SNR into
account; lot of room for improving current scanners.
Fernando
From: "David J. Littleboy" davidjl@gol.com
Newsgroups: rec.photo.equipment.large-format,comp.periphs.scanners
Subject: Re: Relation of bit depth to dynamic range
Date: Tue, 15 Jul 2003
"Ralph Barker" rbarker@pacbell.net wrote:
> Paul, Leonard,
>
> Although late to the discussion, I agree with what I think the two of
> you have agreed to - that a scanner's D-Max is better thought of in
> terms of the scanner's ability to "see" transmission densities, and that
> really has nothing to do with the bit depth of the output.
>
> I think it is better to think of bit depth as controlling the palette of
> shades possible, and, of course, affecting the size of the file. The key
> point, I think, is that at either end of the scale, whether 8-bit (not
> to be confused with 8 bits per channel) or 16-bit, are still white and
> black.
Hmm. I disagree. (Here I assume density values run from 0 = black, 1 =
white).
Each value in the output corresponds to a range of densities in the film.
For values in the middle of the scale, it makes sense to think of a value,
say k, as corresponding to densities in the range k- < error > /2 to
k+ < error > /2, where < error > will be 1/2^n, or the width of each range.
Thus the "meaning" of a value reported by a scanner is the midpoint of the
range that value corresponds to.
When k is 0 (or 2^n-1), that range will be still be k- < error > /2 to
k+ < error > /2. Which means that the center of the range of densities
corresponding to the value 0 is not density 0, but density 0+( < error > /2).
Thus the "meaning" of the value "0" reported by a scanner is the density
1/(2^(n+1)).
Increasing the number of bits moves the center of the range of densities for
which the scanner reports "0" closer to an actual density of 0.
That's what increasing DMax corresponds to: getting 0+(< error > /2) and
(2 ^ n-1)-( < error > /2) closer to the endpoints.
> Thus, it would be just as easy to argue that a D-Max of 9 could
> be "represented" in a single bit - either 0 or 1. But, as Paul points
> out, the resulting image wouldn't be very interesting, detail-wise. ;-)
For a 1-bit scanner, the scanner would report all densities in the range 0
to 0.5 as "0" and all densities in the range 0.5 to 1.0 as "1". So a
reported value of zero should be seen as a "density" of 1/4, and a reported
value of "1" should be seen as a density of 3/4.
David J. Littleboy
davidjl@gol.com
Tokyo, Japan
From: "Bart van der Wolf" bvdwolf@nospam.nl
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: MF Scanner redux- epson 3200 - bah!
Date: Fri, 1 Aug 2003
"David J. Littleboy" davidjl@gol.com wrote
SNIP
> I hate to be a spoil-sport after such an energetic and enthusiastic note,
> but the 3200 is a 1600 dpi scanner.
Owning a 2450, and assuming a similar principle was applied to the 3200
sensor, I have to comment. Technically the above is correct, PER SCAN LINE!
However, because 2 half pixel offset scanlines are used, the sampling pitch
is 3200 ppi. Each color is sampled at 3200 different positions per color.
The fact that the sample areas (probably) overlap, improves the MTF with a
possible small reduction of resolution.
SNIP
> I really doubt that you'll see contrast worth writing home about at anything
> much over 22 lp/mm.
The 2450 (staggered 2400 ppi) resolves a little more than 30 lp/mm of the 47
cycles/mm Nyquist limit. This would lead me to assume that the 3200 scanner
would actually be capable of something like close to 40 lp/mm of it's 63
cy/mm Nyquist limit (but I haven't verified that personally).
Please note that the measured Linepairs were of a resolution target on film,
so of the combined camera lens + film + scanner system. The Nyquist limits
mentioned are for the scanner only.
> Maybe I should try it on my 2450 and we can trade some slides and see if
> it's the sensor or the optics that's the limit...
That is the real thing that matters, but the theory helps to understand. In
fact you should print both scans to the same size, which will reveal that
because the 3200 file needs less enlargement to reach a certain output size,
it will probably improve the output, even if the file resolution was the
same.
Bart
From: "David J. Littleboy" davidjl@gol.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: MF Scanner redux- epson 3200 - bah!
Date: Fri, 1 Aug 2003
"William D. Tallman" wtallman@olypen.com wrote:
> 3200 dpi is not 4000dpi and certainly not 5400 dpi, but it's a whole lot
> better than an effective 1500-2000 dpi. Other than the obvious caveats
> about such mods, any thoughts about what they might produce? A real
> multi-format 3200dpi scanner for $400US is an interesting thought!!!
I hate to be a spoil-sport after such an energetic and enthusiastic note,
but the 3200 is a 1600 dpi scanner.
It's CCD is two 1600 dpi CCDs offset slightly from each other, so what it's
doing is "overstepping" a 1600 dpi sampling system in both directions: in
one by stepping mechanically and in the other by using two offset samples.
The result is _theoretically_ a slight improvement over 1600 dpi, but the
problem with slight theoretical improvements is that for real imaging you
need contrast that's more on the order of 70% than on the order of 10%, and
I really doubt that you'll see contrast worth writing home about at anything
much over 22 lp/mm.
Of course, if you did succeed in upping the contrast at up to 22 lp/mm, you
might be happier with the scans, so it's _not_ a complete waste of time.
Maybe I should try it on my 2450 and we can trade some slides and see if
it's the sensor or the optics that's the limit...
(The 22 lp/mm estimate comes from 70% of the Nyquist frequency for 1600 dpi
sampling and corresponds to using 3 pixels to represent a line pair*. I'd
bet a scanning system that had 80% contrast at 22 lp/mm and rapid falloff
from there wouldn't be all that bad. 7x enlargements would look great on an
Epson 2200.)
*: http://www.clarkvision.com/imagedetail/sampling1.html shows that you need
6, not 3, pixels per line pair to actually render a striped pattern, but the
pattern there is a square wave, not a sine wave. Of course, the features
that we want to render are sharp edges, so we really need 6 pixels. Sigh.
David J. Littleboy
Tokyo, Japan
From: Paul Butzi usenetp@butzi.net
Newsgroups: rec.photo.equipment.large-format,comp.periphs.scanners
Subject: Re: Relation of bit depth to dynamic range
Date: Tue, 08 Jul 2003
Wayne Fulton Fulton@ScanTips_*N0spam*.Com wrote:
>usenetp@butzi.net says...
>>all the way up to 255 representing a density of 255.1
>
>Sorry, I dont get it. I usually tune out at this point, but you said
>"density", and so if log(intensity) = 255 in any fashion, that sure is an
>amazing scanner - my calculator only says Error there. One also wonders of
>course what you might be scanning.
No, I don't really expect to ever represent density values ranging
from 0.1 to 255.1 - I just used those values to point out that the
mapping does not need to have a slope of 1 and an offset of zero.
The point is not that you can use 256 values to represent impossibly
large ranges, the point is that you can use 256 values to represent
any range you please, and that scanners certainly do so.
>
>If you did want to scale the intensity values, then of course I wonder how
>you relay the accurate intensity values to Photoshop? The data doesnt, and
>if you cant and dont, then what's the point?
That's my point. The scanner *does* scale the values, which is why
the pixel values in my image range from zero to 65536 even though the
density range of the negative I scanned runs from Dmin of about .15 to
about 1.8. If we assume your explanation is correct, and that a pixel
value of 65536 would match a negative density of 4.8, when scanning
that negative, I'd never see a pixel value larger than 10^1.8 = 63 and
I can assure you that I most definitely do.
What the scanner does (or at least the combination of the scanner
hardware and the software that drives it) is contrive to scale the
scanner output so that photoshop is delivered values which range from
zero to 65536, where zero is used to represent one end of the range of
densities and 65536 is used to represent the other end of the range.
That the scanner/software combination does this is obvious - if the
slope was fixed at one as you suggest, then the highest values for
pixels would represent the dark areas of the negative/transparency,
and the smallest values would represent the light areas. That's the
opposite of what you'll find if you scan a transparency and fire up
photoshop.
Even worse, if it worked the way you suggest then a scanner output of
1 would correspond to a density of zero, an output of 2 would
correspond to a density of 0.301 (because log10(2) = 0.301), an output
value of 3 would correspond to a density of 0.477, etc. That would
mean that about half of your image would consist of either the value 2
or 3, and we know that doesn't happen, either. Instead the values
from zero to 65536 are spread out linearly in density, not in
illumination. That is, the scanner measures the intensity, uses a log
amplifier or the equivalent to take the log of the intensity and scale
it to the input range of the A/D converter, and the converter converts
that analog signal to the digital representation.
-Paul
--
http://www.butzi.net
From: Paul Butzi usenetp@butzi.net
Newsgroups: rec.photo.equipment.large-format,comp.periphs.scanners
Subject: Re: Relation of bit depth to dynamic range
Date: Tue, 08 Jul 2003
Wayne Fulton Fulton@ScanTips_*N0spam*.Com wrote:
>Yes, dynamic range is about DMax - DMin, and it is not about bit count.
>
>But you still need enough bits to store and hold that result, which is the
>extent that bit depth is involved.
>
>As you said, 8 bits can store numbers representing a max density range of
>up to 2.4 and 16 bits can store numbers representing a max density range
>of up to 4.8. This 4.8 concept is only about number theory, what 16 bits
>theoretically could hold, but is not about what scanners can do.
>
>8 bits can only store numbers from 0 to 255, and thus can only hold
>density range of log(255) - 0 = 2.4 at most.
Huh? I don't follow this at all.
Let's say I'd like to represent a range like this: Dmin = 0.1 Dmax =
255.1
This is a much longer range than your 4.8, but I can happily represent
it using only eight bits by simply using a base density of .1 and a
step of 1.0 so that a zero represents Dmin = 0.1, and a 1 represents a
density of 1.1, a 2 represents a density of 2.1, etc. all the way up
to 255 representing a density of 255.1
The values we use to represent the possible densities when we digitize
conform to the equation
digitized value = (slope) * (measured density) + (constant offset)
(actually, it's probable that it's a bit more flexible than this given
that you can actually account for non-linearity in the detector, etc.
but as a gross generalization for discussion a linear mapping of
domain onto range is good enough)
There's nothing particularly magic about having the slope be 1 and the
constant offset be zero, and I'd be darn surprised if scanners were
actually built that way. In fact, I expect that decent scanners
adjust both the slope and the offset when you fiddle with the control
software and set the 'white point' and the 'contrast' of the scan. If
the slope and offset of the scan are fixed, why would there be any
controls at all in the scanner software? I suppose it would be
possible for the raw output of the scanner to be produced using a
fixed slope and offset, and then the software rescales it, but that
would be what I would call a Very Bad Design.
Even if the slope and offset are fixed and the scaling of values is
done in software, I would be very, very surprised if the slope was 1.
It would make much more sense to adjust the hardware slope to be equal
to the actual Dmax of the sensor/lightsource combo divided by
log(2^bit depth) so that you get maximum possible resolution in the
output. There's no point in wasting output range on densities that
the sensor cannot read, and a density of 4.8 is more or less the
density of a 1" pine board and would probably require a nuclear light
source to get photons through it. Ok, that's an exaggeration but the
point is still valid.
-Paul
--
http://www.butzi.net
Newsgroups: rec.photo.equipment.large-format,comp.periphs.scanners
Subject: Re: Relation of bit depth to dynamic range
From: Wayne Fulton Fulton@ScanTips_*N0spam*.Com
Date: Tue, 08 Jul 2003
len@math.northwestern.edu says...
>
>I thought I understood the relation of bit depth to dynamic range, but
>now I'm not sure. In the latest issue of View Camera, George DeWolfe
>says that the maximum possible dmax for an 8 bit gray scale is 2.4 while
> the maximum possible dmax of a 16 bit gray scale is 4.8. I understand
>mathematically where these figures come from.
>
>log_10(2^8) = 2.41 and log_10(2^16) = 4.82.
>
>But I don't see what this has to do with dmax. The light detecting
>devices in scanners are physical devices capable of responding to some
>range of light intensities. Below a certain level, say I_min, nothing
>will be recorded. Above a certain level, say I_max, additional light
>won't produce any more output. So it seems to me the total range of
>densities the device can handle should be log_10(I_max/I_min). It seems
>to me that the bit depth just determines how finely that range is
>subdivided. For 8 bit, it will be subdivided into 256 distinct levels,
>while for 16 bit, it will be subdivided into 65536 distinct levels. Of
>course, if there is some minimal ratio of intensities which is
>detectable and we assume the scanner is keys to seperating values
>reflected by that minimal ratio, then the two calculations above would
>be relevant. But why can't a scanner with 8 bit depth just use a
>larger step size. After all, the theory behind all this is that if you
>take the human visual system as a standard, then when viewing a single
>gray scale from deepest black to whitest white, a discrete set of 256
>values separating that range will appear continuous.
>
>I would appreciate any comments from experts about just what is going on.
>
>Let me also ask a related question. Just what does my Epson 3200
>scanner and its software do when I use it to scan a b/w negative whose
>maximum density is about 1.4 at 16 bit depth? Does it try to divide
>up the range from 0 to 1.4 into 65536 values, or does it just use
>1.4/3.4 of that range for about 27000 distinct values. What I get using
>Vuescan is values ranging from 0 to 255, and although those may be
>further split up, when I read them into my photoeditor, they also stay
>in that range.
Yes, dynamic range is about DMax - DMin, and it is not about bit count.
But you still need enough bits to store and hold that result, which is the
extent that bit depth is involved.
As you said, 8 bits can store numbers representing a max density range of
up to 2.4 and 16 bits can store numbers representing a max density range
of up to 4.8. This 4.8 concept is only about number theory, what 16 bits
theoretically could hold, but is not about what scanners can do.
8 bits can only store numbers from 0 to 255, and thus can only hold
density range of log(255) - 0 = 2.4 at most. This is not a problem for our
monitors or printers, it is optimistically all the range they have too.
But it helps to start with more range, to hit that narrow target easier.
The reason for more bits is for when you shift the data around over wide
ranges, for example, gamma or histogram or curves. More bits retain more
uniquely different values, instead of combining values into fewer steps.
For a silly but clear example, if somehow it was necessary to shift all of
the 8 bit data up halfway (add 128 to all values), 255 unique data values
would be reduced to only about 128 unique values by my definition, the
rest unfilled with zero. This is a substantial loss, of tones and
contrast. This loss is avoided if you do this shift on 16 bit data
(greatly more than 255 possible unique values remain even if you add 32K
to all values), and then convert the remaining used range to 8 bits.
Either way, you have discarded the bright half of the theoretical range
(if there were infact any data tones actually present there... maybe we
wouldnt do this if there were), but you could still have 256 unique values
remaining then, and can output 256 values for more 8 bit contrast.
No way we can store a 3.4 range in 8 bits of steps which can hold at most
numbers that can represent 2.4. People do argue that fewer bits could
represent a huge range, if it were scaled as necessary to make it fit that
way, but that's actually nonsense. You could perhaps scale it that way,
but there is no way to recover that information from the insufficient
bits. You might put it into 8 bits, but you cannot get it out, so it seems
pointless as a planned goal. All you have to work with is 8 bits now,
regardless what you might have had before, and 8 bits only holds 2.4.
The actual problem is the opposite - 16 bit D/A chips are inexpensive
today, but good CCD chips are not.
Scanners cannot capture 4.8 Dmax of course, nor even 3.8, but computers do
work best in multiples of bytes, like 8 bit or 16 bit words, so it is very
convenient to store say 12 bit data in 16 bit words, simply for better
computer efficiency - vastly faster for manipulation of millions of
pixels. This is where the 16 bit number comes from. The notion of real
16 bit range is fantasy. Scanner ratings of 4.8 are nonsense.
There are 2^^16 =65K steps in 16 bit words, those are the theoretical
values possible, but the real world data from the scanner will not nearly
fill that range from 0 to 65K. The scanner does not have that range, and
the real data will always have a range more like 3.0 than 4.0.
Same concept as pouring a quart of liquid into a gallon jug.
It is a gallon container, but it has a quart in it.
--
Wayne
http://www.scantips.com "A few scanning tips"
From: ralph fuerbringer [rof@mac.com]
Sent: Thu 8/21/2003
To: panorama-l@sci.monash.edu.au
Subject: Re: Noblex user question
the epson 3200 scanner at $400 does up to 4x10 trans or negs ,
extremely well i have 617 though i mostly use the 612 i make with
35,45,55 apo-grandagons and shift. examples of the scans are on
http://homepage.mac.com/rof there is terrific review of the
scanner. go to goggle, type epson 3200 scanner and an english forum has
a 30 page review and analysis of the scanner. selected scanner of the
year in europe.
Brian Walton wrote:
> Scanning 120/220 film is a pain. Is a flat bed better than a drum? I
> would love to use the larger format up to 6X17 but scanning is a
> problem.
>
> Who has been down this path and succeeded?
>
> Brian
From: Peter Marshall [petermarshall@cix.compulink.co.uk]
Sent: Wed 11/19/2003
To: panorama-l@sci.monash.edu.au
Subject: Re: Pakon 35mm Film Scanner
In a rather different price range, the Minolta Dimage Scan MultiPro will
actually scan XPan negs (and others up to around 84mm) although not with
the software supplied. Using Vuescan and a 35mm neg carrier with some of
the bars between frames cut out, you can scan at 4800 dpi.
Of course this is essentially a medium format scanner, which doubtless
accounts for the greater movement even in 35mm mode.
Peter Marshall
Photography Guide at About http://photography.about.com/
email: photography.guide@about.com
From: ralph fuerbringer [rof@mac.com]
Sent: Wed 11/19/2003
To: panorama-l@sci.monash.edu.au
Subject: Re: Pakon 35mm Film Scanner
at the present time Amazon is selling the 3200 Perfection Scanner for
about $250 when you apply their own $100 rebate. It has a 4x9 inch
scan area . The UK super forum review (just put the scanner in google
and it will come up) compares it with a dedicated Nikon 2700 film
scanner at the same level ( it has 5 more) and it is better. Almost as
good as an Imacon--i cant see the diff. I have it and used it to scan
my 6x12 wide angle negs at http://homepage.mac.com/rof
...
From: "MikeWhy" mikewhy@my-deja.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: are digicam users becoming serious amateur photogs - or NOT?
Date: Mon, 29 Dec 2003
"David J. Littleboy" davidjl@gol.com wrote
> MF has real problems: 645 scanned with an Epson 2450 or 3200 isn't as good
> as 6MP dSLRs.
No problems with 2400 dpi scans using the 3200, and B&W; is good for more.
That puts 35mm about on par with 6 MP; 645 is at least better than the 1Ds
or 14n, even if a bit noisier. For me personally, the difference is
meaningful. I downsample the scans to print 13x19, where 6 MP dslr would be
upsampling.
From: "Neil Gould" neil@myplaceofwork.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Fri, 26 Dec 2003
Recently, Bob Monaghan posted:
> I doubt there will be much of an improvement past 8000 dpi with
> current films, though it is interesting to speculate if FUJI and
> others may come out with films enhanced for scanning and sharpening
> rather than standard chemical printing? ;-)
Now, there's an idea worth exploring. I've long suspected that many of the
artifacts that accompany scanning of film is the poorly matched objectives
of each media. The scanners we've been discussing are best suited to
capturing 2D images for printing on 2D media, while film and chemical
prints are both 3D media. It would be intriguing to see if a "2D film"
could be created that would have better properties for scanning.
Regards, and wishing all a safe and happy holiday season,
Neil
From: "David J. Littleboy" davidjl@gol.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Fri, 26 Dec 2003
"MikeWhy" mikewhy@my-deja.com wrote:
> "David J. Littleboy" davidjl@gol.com wrote:
> > You certainly don't have 12 bits of valid data in each channel in the
> > demosaiced result, but you'll have more than 8, leaving the tonality a
> > full order of magnitude better than that of high-res scanned film...
>
> Scanners capture only 9 bits? (2^12 / 10 ==> 4096/10 = 410 ==> 512 = 2^9).
No, that was about _dSLRs_, which I claim (incorrectly?) capture more than 8
bits per color at ISO 100.
IMHO, scanners only capture 5 bits of valid data per color.
When I look at the values of adjacent pixels in areas that I think should be
changing smoothly, I see variations in values of the individual RGB channels
with magnitudes on the order of 3 bits out of 8. That means, to me, that
only the upper 5 bits in the 8 bit value is valid data. These variations
look like noise to me. If you see those variations as noise, that means that
each pixel has only 15 bits of real information. This is controvertial. Some
people think scans have 14 x 3 bits of valid data in each pixel.
> Not what I wanted to hear about the state of the art in scanner technology.
> It would explain why scans feel a little "brittle" when you manipulate them.
>
> Are you sure about this? What's the point of scanning 16-bit?
Getting detail out of the shadows, is what people say. (As Kodak points out,
the apparent graininess of films is worst in the middle of the range: the
noise I'm talking about is in the mid tones.)
http://www.kodak.com/US/en/motion/students/handbook/sensitometric6.jhtml
Note that the magnitude of this noise varies with scanner and scan
resolution in some non-obvious ways, so it's also controversial what portion
of the magnitude of this noise is due to film and what to the fact that you
are scanning film.
David J. Littleboy
Tokyo, Japan
From: "Bart van der Wolf" bvdwolf@no.spam
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Scanner noise
Date: Thu, 25 Dec 2003
"MikeWhy" mikewhy@my-deja.com wrote ..
SNIP
> I bought NeatImage after a brief demo (Thanks, Bill.) Zapped the speckles
> right out of small test snips, although it did cost quite a bit in
> sharpness.
I've been using NeatImage for quite a while now, and it is possible to
have an almost identical sharpness but without the noise. However, it does
require some experience in tweaking the parameters. An important one for
me is to hold back a bit on the removal of luminance noise, which I often
set between 40 and 50% (based on a good noise profile). This will not
eliminate all of the noise (which would look unnatural) but reduce it to
non-distracting proportions.
You can choose to add a little high frequency sharpness in NI, or do it
with a photoeditor. Doing it in the photo editor also allows working with
edge masks for sharpening.
Another very useful technique is to layer the original and a "Neated"
version, while masking and or blending selective parts of the image. This
will allow leaving noise as pseudo detail in very 'busy' parts like tree
branches or rock, and smoothing it in e.g. the sky. That also works
wonders with portraits, smoothing the skin, but leaving the hair 'detail'.
Enjoy this "must have" application for all types of digitized images.
Bart
From: "Bart van der Wolf" bvdwolf@no.spam
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Thu, 25 Dec 2003
"Bob Monaghan" rmonagha@engr.smu.edu wrote
SNIP
> it is also hard for you to explain why an image scanned at 5400 dpi vs.
> 4000 dpi should look better, and many say it does (cf Bart..), if there is
> NO info on the film past 10 MP and it is all noise?
Trying to avoid getting personal with anyone, it is a public forum, the
above observation was for me proof enough that it is possible to have more
detail in a film than a 4000ppi scanner can extract.
Quantifying the information content in a simple number of pixels,
disregards the quality of the pixel value, and is IMHO not meaningful. All
it does is describe is the Nyquist limit to reliable image reconstruction.
Several people have come to similar conclusions, that the probable
practical limit to information content extractable from normal pictorial
film is somewhere between 6000 and 8000 ppi (with diminishing returns).
Scanning at higher ppi sampling pitch still makes sense if one tries to
avoid mathematical interpolation when large output is the goal. It's
arguably better to sample than interpolate between 'known' samples.
See e.g. http://clarkvision.com/imagedetail/ , and in particular:
http://clarkvision.com/imagedetail/scandetail.html and
http://www.clarkvision.com/imagedetail/film.vs.6mpxl.digital.html .
Also noteworthy is the Norman Koren site at:
http://www.normankoren.com/Tutorials/MTF.html , and for his conclusion
based on a calculation model:
http://www.normankoren.com/Tutorials/Scan8000.html (my practical test with
5400 ppi gives somewhat higher values than he calculates, but with
different film and lens).
Bart
From: rmonagha@engr.smu.edu (Bob Monaghan)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: 25 Dec 2003
Hi Bart,
thanks for the URL, esp. last one, he cites very similar levels (4000 dpi
32.7 lpmm for 50% contrast) to David L's estimates...
So I suspect David and rafe are right 95% of the time, i.e., for
most folks using autofocus cameras or handheld or similar resolution
limiting techniques, the scans beyond 4000 dpi are NOT likely to produce
much better results. Here the system resolution is being limited more by
the on-film resolution achieved than by the scanner.
On the other hand, as your experience and Koren's analysis suggests, those
with reasonably high quality images on film may expect to see some
additional image quality from higher end scanners at 5400 dpi and even
8000 dpi, esp. with sharpening effects (to 100+ lpmm from velvia equiv.).
I doubt there will be much of an improvement past 8000 dpi with current
films, though it is interesting to speculate if FUJI and others may come
out with films enhanced for scanning and sharpening rather than standard
chemical printing? ;-)
happy holidays - regards to all - bobm
From: Mxsmanic mxsmanic@hotmail.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Thu, 25 Dec 2003
Lassi HippelSinen writes:
> One-colour pixels sampled with 8 bits?
I'm just averaging it out: 11.4 megabytes of raw data for 11.1
megapixels. That's slightly over eight bits of information per pixel,
total. But since three primary colors need to be represented in an RGB
image and a matrix filter is used for this, that works out to less than
three bits of color information per color, per pixel. Not very much.
Given the above, it is a _mathematical impossibility_ for a 1Ds image to
contain more than 11.4 megabytes of image information. Film beats this
feature very easily even in the most ordinary scans.
> So the raw image hasn't even interpolated the colours from
> the Bayer pattern. Really raw.
Yes. And, as a result, it's a very good indication of the actual amount
of useful image information coming off the sensor. If the raw image is
11.4 megabytes, there is no way that any more than that can ever be
obtained from the camera, no matter how big the normal output TIFF or
JPEG files might be.
This relatively small number also illustrates the huge amount of
information that is sacrificed by a matrix color filter over the sensor.
Fully 2/3 of the image information is lost.
From: "David J. Littleboy" davidjl@gol.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Thu, 25 Dec 2003
"Lassi HippelSinen" lassi.hippelainen@welho.com wrote
> Mxsmanic wrote:
> >... If a 1Ds produces a 11.4 megabyte file, then, the information
> > in the image cannot exceed 11.4 megabytes. That's not even three bits
> > per color, per pixel. There is no way around this.
>
> One-color pixels sampled with 8 bits?
Sampled with 12 bits.
> So the raw image hasn't even
> interpolated the colors from the Bayer pattern. Really raw.
Yes, the dSLRs can all report the actual 12-bit measurements from the sensor
and let you process it however you want, including producing 48-bit tiff
files that have noise that's a lot lower than the grain noise in scanner
images.
You certainly don't have 12 bits of valid data in each channel in the
demosaiced result, but you'll have more than 8, leaving the tonality a full
order of magnitude better than that of high-res scanned film...
(Some people think that the reason dSLR images look so much better that
naive bobm-style resolution measurement comparisons would lead one to
believe is that the perceptual quality is proportional to the product of the
resolution and the inverse of the noise (i.e. total information content). I
personally think that it's simply that film's MTF at the "resolved"
frequencies is too low to contribute to perceived image quality, so the
resolution measurements are meaningless.)
David J. Littleboy
Tokyo, Japan
From: Bruce Graham jbgraham@nowhere.com.au
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Nyquist vs scanners?
Date: Fri, 26 Dec 2003
rmonagha@engr.smu.edu
says...
>
> Hi David,
>
> most interesting, thanks for the insight! David L. has noted he gets
> improvements using a glass carrier to flatten his film, and has a focus
> variation control to adjust optimal focus too (where rafe and others have
> either the AF or fixed focus setups). Your note helps confirm my suspicion
> that film flatness may be a major reason why scanners rarely exceed 50+
> lpmm in practice - fine for most images which probably don't do much
> better either (autofocus or other limitations), but a problem with high
> resolution slides, where even modest 0.1mm mis-focus can mean loss of
> 20-30+ lpmm potential resolution.
>
> film flatness is a big issue in medium format too ;-(
>
> regards bobm
Don't generalise too much Bob. There are basic design differences
between scanners on the market today, with some different characteristics
as a result.
The Canon is reported to have a relatively small aperture lens which is
presumably less affected by depth of field issues. The Nikon has a
larger aperture lens to allow the use of LED light sources which can be
switched at high speed, allowing a single pass RGB and infra-red scan.
There has been a lot of debate on the Nikon depth of field issue and my
reading of that debate is that with care, it is not an issue limiting
scanner performance. By contrast, the Canon (and I think Minolta?) uses
a cheaper cold cathode illuminated pass to get RGB and a separate pass to
get the infrared channel that introduces at best a doubling of scan time
and at worst some registration problems between passes. Given the more
diffused illumination on the Canon/Minolta, there may be a reduction in
resolution compared to the Nikon also (and maybe a reduction of grain
aliasing).
Now I just don't know enough about these issues to draw real conclusions,
but I think I know enough to know the answer will be more complicated
than you are suggesting. (even ignoring other important specs such as
dynamic range, scan speed etc.)
Ed Hamrick, who has played with more scanners than most, as author of a
scanning program that supports almost all the mainstream scanners,
recommends the Nikons on his website for both mid range and high end
scanners, so I doubt he believes there is a significant film flatness
issue with the Nikons.
Hopefully others can shed some light. About all I have learned in about
two years of scanning is there are LOTS of ways to screw up.
Bruce Graham
From: David Gay dgay@lagaffe.CS.Berkeley.EDU
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Nyquist vs scanners?
Date: 24 Dec 2003
rmonagha@engr.smu.edu (Bob Monaghan) writes:
> My best guess now is that those seeing only 35-40+ lpmm from provia either
> have subjects with no high contrast areas, or more likely are being
> limited by various factors to achieving perhaps 40 or 50 lpmm max. on
> provia 100 - not by the film, but by issues like using autofocus step
> focusing, where small focusing errors can "cost" us up to 50% or more of
> potential high resolution on film.
Don't forget the scanner's autofocus, depth-of-field of the scanner optics,
and slide/negative holders that don't hold the film flat enough (with an
LS-2000 I occasionally had to scan the same slide twice, once focusing at
the centre, once at the edges, and then combined the images...).
--
David Gay
dgay@acm.org
From: rmonagha@engr.smu.edu (Bob Monaghan)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Followup-To: rec.photo.equipment.medium-format
Date: 24 Dec 2003
Hi David,
I do agree those numbers are excessive, but they are yours, not mine ;-)
I have consistently suggested that provia can take 10-12X enlargements,
maybe a bit more, but don't recall claiming 20X for it.
the problem with using a 6 MP on 8x10" is that most of the image data is
interpolated and smoothed on the print, and more so as you enlarge more.
the film image isn't interpolating or creating data when enlarged
chemically, it just works with the data it recorded. So these are
different processes.
if you enlarge a 35mm 100 ISO film image by 12X, and a 35mm full format
sensor (6MP?) to 12X, which will have more crisp details and higher
contrast in fine details (hair etc.)? If you do it right, the answer
should be the film, because its resolution limits (100+ lpmm) enable it to
put 70-80 lpmm+ on the film with good technique, while a digicam has a
nyquist limit due to sensor size that limits it to ~50 lpmm.
From other work with 50 lpmm limited images (low $$ zooms etc.) versus
fixed lenses achieving 70-80 lpmm, we know we should see better resolution
and finer contrast in the high resolution images. You should see the same
kinds of differences in film vs. digital, and if you don't, there should
be a reason.
What if that reason is that many scanners are less than optimally aligned,
or that the film isn't being held flat enough (hence, that's why you spent
$ on a scanner that could be adjusted, with a glass carrier to flatten
film and get better results. Presumably, this would be good news to others
like rafe who have fixed scanner optics and want to get more out of them.
the other answer to your complaint is that many pros, who spent $100+ on
high end laser-photomultiplier scans and laser output prints of large
size, do so because it provides much more of the high quality data that is
present on film. They don't do it because they like wasting money, yes?
;-)
I don't expect folks using prosumer scanners to compete with high end
commercial scanners costing 40 to 100+ times as much. But we have already
seen that some folks are able to do rather better than others in image
quality. In your case, David, that's cuz you have that glass carrier etc.,
and others don't, and a focus-able scanner etc. as just one reason
example. Others have found they can recover high resolution info on their
scanners (Bart, Bruce etc.), so calling into questions claims that this
can't be done on similar scanners (e.g., dimage 5400 series).
the difference between 10 MP, which you are seeing with 35mm and others,
and 40 MP, which I claim is there per Kodak's specs (is Kodak's scientists
"insane" for making such claims?) ;-) - that is the difference between 35
or 40 lpmm (which you are seeing at 10 MP per your notes) and 70-80 lpmm,
with some folks seeing 76 lpmm in their dimage 5400 scans.
it is also hard for you to explain why an image scanned at 5400 dpi vs.
4000 dpi should look better, and many say it does (cf Bart..), if there is
NO info on the film past 10 MP and it is all noise? Since the image looks
better, it has to be that going from 62 lpmm (at 4000 dpi per Bart's
posting, IIRC) to 76 lpmm (at dimage 5400 limits) is recovering additional
info from film. So there must be still recoverable info past 60 lpmm up to
76 lpmm, and perhaps even more with a better higher $$ scanner?
In short, claims that film 100 ISO only delivers 10 MP doesn't hold
against the specs of kodak, pop photos lab tests, calculations presented
by me above, and results reported by some scanners (Bart, Bruce..). We
have to conclude that film 100 ISO can sustain up to 100+ lpmm (140 lpmm
at high contrast for provia 100 per specs) and that good technique can
deliver up to 40 MP or so on fine grained ISO 100 films, as mfgers claim.
regards bobm
From: "Neil Gould" neil@myplaceofwork.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Wed, 24 Dec 2003
Recently, David J. Littleboy davidjl@gol.com posted:
> "Neil Gould" neil@myplaceofwork.com wrote:
>> Recently, David J. Littleboy davidjl@gol.com posted:
>>> "Bob Monaghan" rmonagha@engr.smu.edu wrote in message
>>>> I do think there is a big range of tonality in film, because film
>>>> is analog.
>>>
>>> Look up dynamic range in a standard engineering text. In the
>>> presence of noise, analog is digital (in the sense that it is only
>>> capable of rendering a fixed number of discernable levels).
>>>
>> I suspect that you're referring to the signal to noise ratio. Dynamic
>> range refers to the distance between the minimum and maximum values
>> that can be recorded.
>
> I think I meant dynamic range: the distance (in either steps or bits)
> between the minimum and maximum values that can be recorded. (If I
> read the fine print correctly, SNR and dynamic range only differ in
> that SNR can be defined at an arbitrary reference level, and dynamic
> range is always defined by the max and min signals. So the SNR may be
> the same as the dynamic range if you define the reference point as
> the max point.)
The _distance_ between the minimum and maximum values that can be recorded
is the "dynamic range". The SNR determines the accuracy (or quality) of
the recorded signal. They're really quite different measurements. The
"arbitrary reference level" that you're referring to is a point _within_
the dynamic range (even if that point is the max level of the DR). As one
can't define a range and a point as being the same thing (a range of zero
is a point by definition), I really can't see how you arrive at the SNR
and DR being the same.
> An analog signal is capable of holding a number of levels that is
> determined by the noise, is what I was trying to say.
However, the number of values that can be represented in an analog signal
is independent of the system noise. If you're referring to the error
introduced by system noise, then we're once again talking about the
behavior mostly at the extremes of the dynamic range. With a camera, one
form of system noise is vibration, and it's clear that this impacts
resolution more than tonal range or accuracy, for example.
>> Noise affects dynamic range most critically at the
>> extremes because recorded values become ambiguous due to limitations
>> of the sensor. This says *nothing* about the "number of discernable
>> levels" within the unambiguous portion (mid range) of the dynamic
>> range. In any analog recording system, that number is close to
>> infinity for all practical purposes.
>
> I think you are confusing scanner and normal engineering use of the
> concept dynamic range. (Actually, I think the scanner types are
> confused, but that's another rant {g}.)
>
Well, my folks paid for me to go to engineering school for a reason, I
expect. 8-)
But, to the point at hand, the dynamic range of a scanner is still the
distance between maximum and minimum values that the sensor(s) can record.
It's specified as a dMax value. Quantization, or the bit depth, is the
ability to resolve small differences to discreet values. While
quantization errors create a deviation from the source signal (image),
they really have no impact on the dynamic range because these errors occur
*throughout* the dynamic range.
>> Two shots of the same still life subject taken sequentially on the
>> same roll will *not* render the same image at the dye cloud level.
>
> I've not explained myself well: that is exactly my point: the
> randomness due to the dye clouds can be seen as noise in the signal
> one is trying to record. (Forget that there's a scanner in here for a
> bit: film measures an image, and that measurement is noisy. The
> scanner then tries to measure that measurement. That's why we get so
> dizzy: there's (at least) two levels getting confused.
It's probably not convenient to call the kind of error introduced by
"random sensor location" noise. The analogy to digital is that the Bayer
sensor artifacts and pixelization caused by geometrically rigid sensor
locations also create deviations from the scene, but in a different way. I
don't think that these deviations are what we are thinking about when we
talk about the "noise" in an image.
> The noise defines the dynamic range: that's what noise means. The
> problem is that the noise is different depending on the resolution
> with which you look at the film. Through an 8x loupe, there's no
> noise. Through a 60x microscope, you see grain noise, the colors
> don't look as bright, and the tonal range is reduced.
Noise restricts the dynamic range, but doesn't define it, as explained
earlier. I'm not sure what you're getting at w/r/t the 60x microscope. The
tonal range of film can't be viewed in that way. You can see grains, but
for one thing, the illumination of those grains is not correct for
reproducing the colors or tonal range of an image. In effect, the errors
that are introduced by viewing film under a microscope is the same kind of
"noise" you are complaining about. I'd rather call it a distortion, as the
signal is clear... it just doesn't represent the subject very well.
>> While the effects of randomized imaging units is technically
>> "noise", it's hardly "worse" than simply limiting the tonality and
>> tossing the information up front (not to mention other artifacts),
>> as happens with digital. It all boils down to what you like to look
>> at. Some of us don't seem to be as impressed by the artifacts of
>> limited tonality as others.
>
> You may be forgetting that I'm (implicitly) comparing regions with
> the same number of pixels* in digital as in the film, since the
> argument that Bobm presented was that film has infinite tonality.
> Obviously, digital has fewer pixels than film. But at the pixel
> level, film, at the sorts of magnifications Bobm is assuming, is very
> ugly stuff, and he's claiming that it has infinite tonality.
I wasn't forgetting it as much as disregarding such comparisons as
irrelevant. Both methods have deviated from the original scene, but in
different ways. While noise is an issue, it's becoming more apparent to me
that what we're really talking about are distortion and quantization
errors.
>> There are differences between optical and digital image reproduction
>> that I've missed here. For example, in an optical process, both the
>> recording medium and the reproduction medium are essentially 3D.
>> This takes advantage of the random nature of grain or dye clouds to
>> expand the tonal range that the viewer can experience.
>
> Film looks grainy as soon as you get into enlargements on the order
> that scanners implement, and I see nothing wrong with seeing that
> grain as noise.
Well... I do. The grain is an inherent component of the recording media.
With film, there are at least two kinds of noise involved; that which is a
deviation from the original scene (quantization error), and that which is
caused by impurities in the film (i.e. spurious conversion). Digital
imaging suffers from these two errors, quantization error being obvious
and impurities in the sensor leading to spurious conversion, and adds
system noise. To prevent this from rendering horrid images, *tonal range
gets tossed*, resulting in the "flattening" of areas of subtle gradation
to a limited range of values. Whether this is offensive to the viewer is
purely subjective.
Bottom line is that the "noise" you're referring to isn't a quality of the
recording media (the film), it's an artifact of the scanning process.
>> To that end, one might think of scanning as a process of artifact
>> management, and is much an art as a science. 8-)
>
> I just go through the roof when faced with near-religious
> arguments about the infinite tonality of film.
I'm not claiming that film has "infinite tonality", or even close. If that
were possible, we'd only need one kind of film. 8-)
I'm saying that, as an analog medium, film has a large array of tonal
values that can be represented. There is a significant difference between
an infinite number of possible values and infinite tonality. Consider the
range between any two adjacent quantization values. By increasing the
number of grains involved, you can present more than two values in that
range, but that hardly approaches "infinite tonality". If nothing is done
to increase the bit depth, you won't be adding any tonal detail by
increasing the xy matrix of digital sensors. I also contend that the
practical range of tonal values that "film" captures exceeds the
quantization capabililties of prosumer CCD scanners. By a significant
margin. But, that, too, is another rant! ;-)
Regards,
Neil
From: "Neil Gould" neil@myplaceofwork.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Wed, 24 Dec 2003
Recently, David J. Littleboy davidjl@gol.com posted:
> "Bob Monaghan" rmonagha@engr.smu.edu wrote in message
>> some interesting points, but we _do_ have some different views,
>> makes it more of a challenge ;-)
>>
>> I do think there is a big range of tonality in film, because film is analog.
>
> Look up dynamic range in a standard engineering text. In the presence
> of noise, analog is digital (in the sense that it is only capable of
> rendering a fixed number of discernable levels).
I suspect that you're referring to the signal to noise ratio. Dynamic
range refers to the distance between the minimum and maximum values that
can be recorded. Noise affects dynamic range most critically at the
extremes because recorded values become ambiguous due to limitations of
the sensor. This says *nothing* about the "number of discernable levels"
within the unambiguous portion (mid range) of the dynamic range. In any
analog recording system, that number is close to infinity for all
practical purposes.
Of course, there are sources of error in the analog recording process;
this is noise relative to the signal, but, such errors are not necessarily
discrete in nature. The efforts to reproduce an analog recording
introduces error, again relative to the signal. Still, there is no
relevance to "fixed numbers"; different analog reproductions will vary in
different ways.
The A/D process adds "fixed number" limitations to the number of
discernable levels in the mid range, but this is largely mandated by the
requirement of binary encoding. So, with digital the noise errors are
irreparably introduced "up front".
> Any time you measure a physical signal, increasing the resolution
> increases the noise. If you see film as a physical signal that uses
> discrete dye clouds to represent a color, it's clear that the
> tonality resolution goes down with area.
Two shots of the same still life subject taken sequentially on the same
roll will *not* render the same image at the dye cloud level. If the dye
clouds were truly discrete, as you posit, the results would be identical.
I challenge you to accomplish this. While one can regard the differences
between these images as being affected by the "noise" of film due to grain
or dye cloud variances, it is related to the signal (i.e. the subject) and
has no impact on the number of discrete levels -- i.e. tonal range -- that
might be represented within each image.
>> When we print those images, we don't usually do so at 55X,
>> however, so the tonality effect in film is inherently superior to
>> digital's limited step bit depth (being analog etc.), and that
>> tonality is quite acceptable to me and others at 10X or 12X
>> enlargements, yes? ;-)
>
> No. The noise in film is worse than the noise in digital.
The noise in "film" is different than the noise in digital. Keep in mind
that we're talking about the accurate reproduction of the original scene,
and errors in that process is the noise. With film, you're looking at the
variances due to the randomization of grain or dye clouds. For the
purposes of this discussion, neither the dynamic range nor the tonality
are impacted. While the effects of randomized imaging units is technically
"noise", it's hardly "worse" than simply limiting the tonality and tossing
the information up front (not to mention other artifacts), as happens with
digital. It all boils down to what you like to look at. Some of us don't
seem to be as impressed by the artifacts of limited tonality as others.
There are differences between optical and digital image reproduction that
I've missed here. For example, in an optical process, both the recording
medium and the reproduction medium are essentially 3D. This takes
advantage of the random nature of grain or dye clouds to expand the tonal
range that the viewer can experience. When scanning (and in a similar way
digital printing), the error introduced by the rendering of this 3D data
into 2D adds considerable noise (e.g. inaccurate representation of the
signal). To that end, one might think of scanning as a process of artifact
management, and is much an art as a science. 8-)
Regards,
Neil
From: Mxsmanic mxsmanic@hotmail.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Wed, 24 Dec 2003
David J. Littleboy writes:
> Not at all. You can save the data as captured as 12 or 14 or 16 bits times
> number of pixels, but after converting to RGB, you can't drop bits without
> losing information.
Since conversion to RGB created a tremendous amount of redundancy,
dropping bits may not lose any information. If you have 12x3600x2400
bits of information straight from the sensor, you have 8.64 megabits of
information at most; conversion to RGB will produce an image with
perhaps 415 megabits of data, but the actual information in the image
will still be only 8.64 megabits. So you can drop nearly 98% of the RGB
image without losing any information at all.
There is nothing in a digital image beyond what is provided in the raw
data. If the raw data amounts to x bits, there is no way that any
conversion thereafter will provide more than x bits of real information,
period.
The 1Ds provides a raw file size of about 11.4 megabytes. That works
out to just eight bits per pixel, or less than three bits per primary
color. It is a _mathematical impossibility_ for an image to contain
more information than the number of bits in the raw image file.
From: "Bart van der Wolf" bvdwolf@no.spam
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: standard techpan slide available etc. Re: scanners - good enough?
Date: Tue, 23 Dec 2003
"Bob Monaghan" rmonagha@engr.smu.edu wrote
> Hi Bart,
>
> Thanks very much for the URL to your site and sharing your efforts !!
> http://www.xs4all.nl/~bvdwolf/main/foto/scan/se5400/se5400.htm
>
> You have been able to demonstrate getting 62+ lpmm from provia 100
> with a LS4000 scanner, and over 76 lpmm with a SE5400 scanner, using high
> contrast test charts. So clearly the scanners _ARE_ capable of doing
> better than 40 or 50 lpmm resolution limits, as you have shown. Good work!
You're welcome.
I have also made available a test target that's better suited for testing
both analog AND digital cameras (optics+sensor combined). There's a
version that can be printed at the indicated resolutions:
http://www.xs4all.nl/~bvdwolf/main/downloads/Jtf60cy-100mm_600ppi.gif for
HP inkjet printers and,
http://www.xs4all.nl/~bvdwolf/main/downloads/Jtf60cy-100mm_720ppi.gif for
Epson printers.
When printed without enhancements on glossy paper it produces a 100x100mm
test target which can be easily used for quantifying the limiting
resolution for a target contrast of say 100:1.
Shooting distance is not critical, something like between 50-100x the
focal length will be adequate. The blur diameter for a given well focused
lens/film combination is always the same, regardless of distance!
> from http://www.fujifilm.com/JSP/fuji/epartners/bin/Provia100f.pdf we see
> is below 30% contrast response at ~60 lpmm and well below that for 75
> lpmm (~20%?). So we can also suggest that proposed MTF limits for adequate
> scanning of 50% or more are not necessary, and that we can use scanners
> down to 20% contrast range, possibly below...
The ISO assumes a 10% modulation to correspond loosely to the limiting
resolution of a bi-tonal bar chart.
> This now suggests that the problems lie, as you inferred, with starting
> with a good high resolution image. This corresponds to my own observations
> that beating 50 lpmm is hard work(see http://medfmt.8k.com/mf/limits.html)
>
> and especially critical wrt focusing - a minor error in focusing can
> easily "cost" us 50% or 60% of resolution, even at f/8, - see
> mf/critical.html top chart on high res film resolution falloff with small
> (2mm) focusing errors and 35mm lenses etc.
Yes. High ppi scanning is merciless in uncovering lack of focus, (motion)
blur, or lens aberrations.
> so we can suspect that most scanners, if tested with a series of high
> resolution film images (including some high contrast test chart scenes for
> calibration), might well be able to double (e.g., David L's 35 lpmm with
> provia 100 vs. your 76 lpmm result) the resolution and contrast results
> from scanning film...
The home made chart proposal above is in my experience rather tolerant to
film contrast differences and thus usually gives approximately the same
outcome regardless the shooting distance. The limiting resolution for this
particular 60 cycle/circumference chart can be calculated as:
(60/pi)/blur_circle_diameter. The blur_circle_diameter can also be
expressed as a number of pixels multiplied by the sampling pitch in
millimeters, which gives lp/mm.
> STANDARD SLIDE AVAILABLE:
>
> finally, this re-emphasizes the need to have a standard very high
> resolution slide for scanning, preferably one with some high contrast
> chart imagery, and made with techpan, currently the king of high
> resolution and high contrast (B&W;) films, and preferably at up to 100 lpmm
Shooting my proposed target on Provia/Velvia should provide a scan or
optical print target that already exceeds 4000ppi scanner resolution (and
probably 5400 ppi as well). Shooting the same target on TechPan would push
the limit even further, but shooting it on whatever film you normally use
would give a very good indication of practical limits (your lens at your
aperture with your film, center or corner).
Bart
From: Bruce Graham jbgraham@nowhere.com.au
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Nyquist limiting scanners to 50 lpmm Re: Scan Comparison Site
Date: Tue, 23 Dec 2003
...
Bob
I'm a bit rusty on this stuff (I learned it 40 years ago) but I think
that Mr Nyquist said you have to sample at double the frequency, not
double and then double the frequency you suggest.
I think the spatial frequency IS the lp/mm ie. the one cycle is a single
line pair. (one repetition of black to white). Thus it is necessary to
sample in excess of double the lp/mm, not double double as you suggest.
4000dpi = 160 samples/mm = 80lp/mm
Now in practice, it is difficult to get really close to the Nyquist limit
without special sharp roll-off filter transfer functions, which I don't
have on my scanner.
What I think I'm seeing with my Canon 4000dpi scanner is sometimes about
60-70 lp/mm and sometimes aliasing from sharp grain edges. I can control
the aliasing when I see it, by manual defocus of the scanner, which is a
slow roll-off filter which probably accounts for the extra factor of two
you quote (40 lp/mm). But that extra factor of two is not from
information theory, just from imperfections in implementation. An anti-
aliasing filter could in theory remove most of your last factor of two.
But as David L. says, usually there is not much point because of the
grain noise.
Bruce Graham
From: "Bart van der Wolf" bvdwolf@no.spam
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Mon, 22 Dec 2003
"David J. Littleboy" davidjl@gol.com wrote
> "Bart van der Wolf" bvdwolf@no.spam wrote:
> > I've seen this mentioned several times, but I was too busy to participate
> > in the thread, until now.
> > I disagree with the observations mentioned, based on personal experience
> > with an LS4000 and an SE5400. Something like a quantifyable 70-80 lp/mm
> > can be extracted by scanning (Provia) film, but it requires a good lens on
> > a tripod mounted or steady handheld exposure to have so much detail in the
> > first place.
>
> http://www.terrapinphoto.com/jmdavis/
The main problem with those "comparisons" is that they don't convey Field
of View. A pixel by pixel comparison is not too meaningful because of
different magnification factors.
> Hmm. I still disagree. I just don't see it in real scans. 70-80 lp/mm
> implies the same sort of image quality Rafe shows in his "perfect scan"
> images. That just doesn't happen in real scans.
http://www.xs4all.nl/~bvdwolf/main/foto/scan/se5400/se5400.htm , example 1
will demonstrate I can pull 76 quantifyable lp/mm from a (admittedly 35mm)
film, and I'm not so vain as to think I'm the only one who can.
> > The only important thing is the combined MTF of lens AND film together
> > (contrast/modulation near the limiting resolution is important).
>
> I suspect that it's the MTF that's the problem here. Someone might look at
> the TechPan scans I put up on that site and jump up and down and shout "75
> lp/mm 75 lp/mm!!!", but that doesn't make those scans anything other than
> useless mush at much over 10x on a print. 75 lp/mm at under 5% MTF just
> isn't useful.
That depends on the rest of the workflow. Properly sharpened at its final
output size it will help.
> > An optical microscope can show that there's more fine detail than claimed
> > by some 4000 ppi scan results.
>
> I look at just about every one of my frames with a 60-100x microscope, and
> I've never seen anything but the same fuzz that the scanner reports.
As I said, it depends on the camera lens and focus and lack of motion blur.
> That includes the Tech Pan. To the best that I can tell, talking about anything
> over 35 to 40 lp/mm on film is head-in-the-clounds ivory-tower fantasy.
I've got some more examples (I have to add them to my website) that will
show the difference between a 4000 ppi scan and a 5400 ppi one from the
same piece of film, so it must be additional resolution (and it's much
higher than the numbers you quote). I can even get more than 30 lp/mm from
my Epson 2450, which isn't exactly the sharpest scanner around, so it's
hard to understand you can't get better scans from a dedicated
filmscanner.
Bart
From: "Bart van der Wolf" bvdwolf@no.spam
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: scanners - good enough?
Date: Mon, 22 Dec 2003
...
SNIP
The only important thing is the combined MTF of lens AND film together
(contrast/modulation near the limiting resolution is important). An
optical microscope can show that there's more fine detail than claimed by
some 4000 ppi scan results. This will be the basis for both optical
enlargement (with unknown lens MTF and focus accuracy) and scanning (with
a good chance on optimal focus due to the collimated nature of most
scanner lighting systems).
Bart
From: stacey fotocord@yahoo.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Fuji's warning, 63% of dig images users at risk? Re: Digital_Film
Date: Thu, 18 Dec 2003
Rafe B. wrote:
> Is there a film scanner in existence that can
> resolve the claimed resolution?
Maybe not but what does that say other than there isn't a film scanner that
can get all the info that is there? If 35LPmm is all film can resolve, why
can anyone see the difference between a $65 35-80 zoom and a $1200 L series
lens given a $65 lens can easily resolve 35LPmm?
--
Stacey
From: stacey fotocord@yahoo.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Scan Comparison Site, Call for Sample Scans
Date: Thu, 18 Dec 2003
Rafe B. wrote:
> I've posted a TIF corresponding ot one of the LS-8000
> JPG scan snippets.
>
> Here's the TIF:
> http://www.terrapinphoto.com/jmdavis/chrome_41_tif.tif
> and here's the JPG:
> http://www.terrapinphoto.com/jmdavis/chrome_41_ice.jpg
> In terms of viewable detail, there's not one iota of
> difference between these two images.
??? You must be blind if you think these look the same! How can you use info
like this as a display of what different scans look like? The tiff looks
like a pretty nice scan, the jpg looks like a noisy mess that would make me
think it's unusable. Maybe that's your point, to -make- film look worse
than your loved digicams? Thanx for showing this so I know how bogus this
info you're posting is..
--
Stacey
From: "David J. Littleboy" davidjl@gol.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Fuji's warning, 63% of dig images users at risk? Re: Digital_Film
Date: Tue, 16 Dec 2003
...
Really. As a practical matter, my 4000 dpi film scans look very much like
what the film looks like under a 60x microscope (Hi, my name's Dave, and I'm
a grain sniffer), and film looks like mush at 60x. 4000 dpi lightly
sharpened, downsampled to 60% (2400 dpi) and lightly sharpened again, looks
roughly as sharp as dSLR originals, maybe not quite as good as the best that
a good dSLR gets with a sharp prime at f/8, but in the ballpark.
That's 7.3 MP from 35mm. That's all folks. And (see below) the noise is a
lot worse than digital originals. A lot worse.
> "Bob Monaghan" rmonagha@engr.smu.edu wrote:
> > okay, let's do the math, alright? ;-)
> > start with 108 lpmm resolution on midspeed film (the max rating for provia
> > 100f film from Kodak is 140 lpmm, so this allows for lens losses etc.).
What a joke. Provia is 35 lp/mm at 50%MTF, and due to the grain noise, 25%
MTF would be essentially useless for practical imaging: detail in textures
can be very low contrast and digital users are finding that high ISO images
(even after noise reduction) just don't have the fine detail.
Reality check: the 1Ds is barely 40 lp/mm, if that. But everyone who has
compared the 1Ds to Provia finds the 1Ds comes out ahead. Significantly
ahead. So if you think Provia is a 100 lp/mm sensor, your approach is very
wrong.
> > we have a film array equiv. to 5184 by 7776, or 40,310,784 - or 40
> > Megapixels. Here each pixel could be 24 bit color, or 48 bits etc.
Film has grain noise. That grain noise is inversely proportional to
resolution. At 4000 dpi, the best films are lucky to have 6 bits of valid
data (dynamic range). ISO 100 dSLR images (RAW capture to 16-bit TIFFs) have
9 or 10 bits of valid data. That's an order of magnitude more dynamic range.
(Downsampling improves dynamic range by a bit. Maybe. So film at 2400 dpi is
only a factor of 4 worse than digital.)
David J. Littleboy
Tokyo, Japan
[Ed. note: thanks to gordon moat for sharing these tips on improving scanner results]
Date: Wed, 31 Dec 2003
From: Gordon Moat moat@attglobal.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Nyquist vs scanners?
Greetings Bob M.,
Bob Monaghan wrote:
> David L. has noted he gets
> improvements using a glass carrier to flatten his film, and has a focus
> variation control to adjust optimal focus too (where rafe and others have
> either the AF or fixed focus setups).
To add a bit more to this, there is also an ability to minimize the
appearance of noise (grain) in the scan by carefully adjusting focus off the
autofocus distance the scanner determines. Sometimes this works slightly in
front of the suggested distance, and other times slightly behind it.
RTponses Photo had an article within the last year about scanning using a
glass mount and drum scanning oil. This was a set-up for use with a film
scanner, and only one drop of oil was used on each side of the glass mount.
The most noticeable benefit would be for scratched film, though there was
also a noticeable improvement in edge definition, shadow detail, and colour
tonality. While not fast, nor easy to do often, there may be some film that
can benefit greatly from this technique.
The other similar technique I have heard this done with was flat scanners and
scanning film. Obviously, the cleaning issues are a negative consideration.
Anyway, the drum oil helps avoid Newton rings.
> Your note helps confirm my suspicion
> that film flatness may be a major reason why scanners rarely exceed 50+
> lpmm in practice - fine for most images which probably don't do much
> better either (autofocus or other limitations), but a problem with high
> resolution slides, where even modest 0.1mm mis-focus can mean loss of
> 20-30+ lpmm potential resolution.
Some films curl more than others, and storage can also make that better or
worse. With mounted transparencies, I think it would be slightly less of an
issue.
> film flatness is a big issue in medium format too ;-(
Zeiss seem to think so, though my though is that many subjects of photography
would not be diminished by the lack of perfect flatness. If it was mainly
architecture to consider, then perhaps the problem might be more apparent.
Aesthetic concerns will often be much more noticeable than technical
deficiencies, except for some gear junkies with nothing better to do with
their time than pick at technicalities.
The evenness of the scan is also what prompted Imacon to make a curved
scanning carrier. While that mimics what drum scanners do, there is no oil
used. I think the drum scanning oil use in more than just drum scanners can
give many more benefits . . . just a little more experimenting needed.
Ciao!
Gordon Moat
Alliance Graphique Studio
http://www.allgstudio.com
From: Mxsmanic mxsmanic@hotmail.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Film scanners -- worth it?
Date: Wed, 31 Dec 2003
Carl Bevil writes:
> Anyone out there have experience with film scanners?
Lots.
> Is this worth getting, or will I be disappointed with
> the results? Do you really need to shell out the big
> bucks to get a film scanner that can do 35mm justice?
You don't have to shell out big bucks, as even an inexpensive scanner
will blow away anything you could ever hope to obtain from scanning
prints.
If you shoot film and you want digital images from that film, your only
realistic choice is a film scanner. It is worth what you pay for it.
Trying to scan prints, or even trying to scan film with a non-film
scanner, throws 95% of the quality of the film image into the
wastebasket. Once you scan film directly with a film scanner, you'll be
amazed at just how nice those film images actually are.
From: Robert Feinman robertdfeinman@netscape.net
Newsgroups: rec.photo.equipment.medium-format
Subject: slightly OT: scanning resolution vs print size
Date: Wed, 31 Dec 2003
After getting differing advice on producing smaller magnification
prints from my scanner I decided to run some experiments.
With all the discussion of the limits of resolution in scanners
and film in this group recently I thought some of you might
be interested in the results.
Just to keep things honest, my panoramas are medium format, at
least in one dimension!
The question I tried to answer is what's the best way to produce
a 4.5x enlargement from the Minolta 5400 original and are such
high resolution scanners even needed for "normal" sized prints.
Follow the tips link on my home page and scroll down to the "new"
tip.
--
Robert D Feinman
robertdfeinman@netscape.net
Landscapes, Cityscapes, Panoramas and Photoshop Tips
http://robertdfeinman.com
From: "pioe[rmv]" "pioe[rmv]"@coldsiberia.org
Newsgroups: rec.photo.equipment.35mm,rec.photo.equipment.medium-format
Subject: Re: Scan Comparison Site, Call for Sample Scans
Date: Mon, 22 Dec 2003
David J. Littleboy wrote:
> Given the small amount of high frequency
> information (other than grain noise) in scans, the enormous tiff file macho
> in scanner users is quite misplaced.
No, this is positively wrong. If you want the highest possible quality
and editability, everything must be recorded and stored in lossless
formats. Not only are you losing detail, JPEG means you are limited to 8
bits per channel.
You are supposed to know this, David, and you should not encourage the
kind of cutting corners it is to use JPEG as a storage format.
But, for comparison purposes, everything that is required is that all
the files compared be in the same file format. The information lost in
the compression will of course be even more insignificant in relation to
the differences between the scanners and cameras in question, and they
will be the same for all. To demand TIFF in a comparison is totally
without justification, and this should not be confused with the entirely
rational statement that master files should be recorded and stored
losslessly.
Per Inge Oestmoen, Norway
From: "Neil Gould" neil@myplaceofwork.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Nyquist limiting scanners to 50 lpmm Re: Scan Comparison Site
Date: Sun, 21 Dec 2003
Recently, David J. Littleboy posted:
> "Bob Monaghan" rmonagha@engr.smu.edu wrote
>>
>> Hi David,
>>
>> The reason film scanners can not achieve more than 40 or 50 lpmm
>> equiv. quality is the Nyquist law, which roughly states that you
>> have to sample at least twice as fast as the highest frequency
>> signal you want to reliably detect. This is a basic law of
>> information theory...
>
> Thank you. I understand that quite well.
>
> What I also understand is that if you make a 13x19 from 35mm, it'll
> look like soft mush (whatever film-to-paper technology you use), and
> if you make a 13x19 from an 11MP 1Ds image, it'll look quite a bit
> nicer. And if you make a 13x19 from 645, it might look a tad nicer
> than the 1Ds. Maybe.
>
What does "nicer" mean? In what dimension?
> This isn't about scanning; it's about film having inadequate
> performance to support 13x enlargements.
>
> Film being inadeqiate to support 13x enlargements shouldn't be news
> to you: it is, in fact, common knowledge that everyone agrees on.
> Even 645 is considered inadequate to make 13x19 prints.
>
> (Aside: Since film has inadequate performace to support 13x
> enargements, 4000 dpi is a nice overkill resolution to scan at since
> profession digital printing uses 300 dpi as standard.)
>
> The point here, is that the various theories you've proposed ("film
> _usefully and meaningfully_ resolves 100 lp/mm, as well as your
> incessant quoting the "ISO 400 35mm is worth 23MP" and "ISO 100 35mm
> film is worth 40MP" insanities from Kodak) are grossly at odds with
> objective reality.
The discussions that I've seen on this topic always presume something
about the image without being clear. When one compares images based on the
size of the data, I suspect that they're only looking at the apparent
"sharpness" (or pseudo-sharpness) of the image, rather than, for example,
the tonal gradation. What is being disregarded is that a 300 dpi image @
36 bits (12 b/p) *will* be quite a bit smaller than that image at 48 bits
(16 b/p). So, it makes no sense to me to make statements about the
"insanities" of the relative data sizes *unless* one refers to *both* the
xy resolution *and* the bit depth of images.
What I'm calling "pseudo-sharpness" is the higher contrast that
accompanies lower bit depths. It makes the transitions between colors
appear "sharper" simply by the inability to represent the original
gradations in an image. Consequently, images that *can* represent the
gradations may appear "softer".
A key question that addresses the relationship between data size and
images is the equivalent bit depth of "film". Perhaps Kodak and others
have a different notion of what it takes than those that think that 11 mp
is more than enough to match 35 mm film? A lot depends on the composition.
So, I really am confused when comments about "nicer" images are made in
lieu of comments about the nature of the image.
Another presumption is that there is some relevance to 300 dpi. This
became a popular notion when creating images for offset presswork about 20
years ago, and later carried over to inkjet technologies. Prior to 20
years ago, the limitation of pre-press to reproduce xy resolution was
based on the availability of separation screens. As scanning technology
and imagesetters replaced mechanical separations, the xy resolution has
continuously increased over time. Therefore, I suspect that these
limitations are matters of convenience based on the notion that if a
printer can't resolve images above 300 dpi, then anything more than that
is irrelevant. Now that it is commonplace for offset presswork to print
beyond 150 lpi (the original basis for 300 dpi resolution based on the
Nyquist theorem), the ideal xy resolution has increased accordingly.
Similarly, as Hexachrome technologies can successfully represent a wider
gamut than CMYK, the greater bit depth of images could be exploited. A 200
lpi offset job is not the exotic dream that it was 20 years ago, and 400
dpi images are the "norm" for that kind of work. Even higher line screens
are used in today's "exotic" presswork. Coupled with Hexachrome, the
requisite data size of an acceptable source file has grown considerably.
As inkjet technology continues to improve, both through the use of higher
xy resolution and multiple ink colors for increasing the bit depth
representation, I expect to see the same kind of increase in the data size
of "acceptable" source files.
Regards,
--
Neil Gould
Terra Tu AV - www.terratu.com
Technical Graphics & Media
From: Peter Irwin pirwin@ktb.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Nyquist vs scanners? Re: How big is the resolution on 35mm?
Date: Sat, 20 Dec 2003
Bob Monaghan rmonagha@engr.smu.edu wrote:
> but this raises an interesting point - what is the Nyquist limit on these
> scanners? Is it really the 4,000 dpi or 78 lpmm claimed?
Sampling theorem says that the maximum bandwidth you can have is half
the sampling rate. 4000 dpi translates to 2000 lp/inch or 78.7 lp/mm.
The performance of the antialiasing filters will always knock it down
somewhat from this, but the Nyquist limit is 2 samples per line pair
-not intuitive I know, but it will work in video the same as it does
in audio if the anti-alias and reconstruction filtering is handled
properly.
One thought on testing scanners: you might try shooting some
Eastman Direct MP film 5360 in a camera. It is slow (EI about
0.4) and orthochromatic, but the resolution of the film is
staggering: 320 lp/mm at 1.6:1 contrast. I've not been able to
get more than half that actually visible on film, but it might
make a very interesting test of a scanner. I can't see any grain
in it at all.
Peter.
--
pirwin@ktb.net
Date: Mon, 15 Dec 2003
From: Gordon Moat moat@attglobal.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How big is the resolution on a normal 35 mm negative?
Paul Repacholi wrote:
> Gordon Moat moat@attglobal.net writes:
>
> > At the beginning of this year, a drum scanner was introduced that
> > can do 64000 (three zeroes!) dpi. It is capable of resolving down to
> > about 3 Ym, nearly the size of film grains. Unfortunately, the
> > resulting files can be quite huge.
>
> Got a URL for it?
Okay, I found the reference:
http://www.aztek.com/Products/Premier.htm
What is really cool is that it can take up to 12" by 12" film. This would
be great for any large format scanning.
Ciao!
Date: Sun, 14 Dec 2003
From: Gordon Moat moat@attglobal.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: easy MTF for digital cameras.. Re: How big is the resolution
Bob Monaghan wrote:
> Hi Gordon,
Hello Bob,
> yes, some very good points. Unfortunately, many cameras internally
> process the data using complex algorithms, doing everything from handling
> chip edge effects to removing aliasing artifacts. Even "raw" data may be
> pretty processed ;-)
Very true. I was just reading about an Imacon development in software that
corrects for MoirT. The idea was developed to eliminate the need for a
physical anti MoirT filter. Kodak has also started
> . . . . . . I agree that the E-6 standardized slide is probably better for a
> standardized scanner test than any random sample, as most mfgers would be
> smart enough to cheat to show their product at its best. [my boss bought a
> big $$ color laser printer for our media lab, largely on the basis of the
> built-in test print image quality, which was stunning. We never equaled
> that quality in practice, of course, because it had been carefully tweaked
> to exactly match the laser printer dynamics for optimal quality ;-) ;-) ]
Your print example reminds me of one of my fellow graduates. He went on to
work for a company that created optimized print samples. These were included
with scanner comparison images, and a few ended up on store shelves as sample
prints directly from printers. Everything from high end to low end, and in
between. Of course, given all day, every working day, almost anyone might get
great printing results.
Anyway, back to the Q-60 target. I really think that is the ideal test
subject, since there is less variability than with any other image.
> you are right that billboard specs are low, and ditto movie theatre, just
> hold an 8x10" out to cover up the movie screen, and see how small the
> screen really is (unless you also sit in the front rows ;-)
Exactly, and the viewing distance to size is the main consideration. The
other issue is the speed of printing such large pieces.
> you can compute the maximum resolution from feature sizes, but factors
> like the anti-aliasing filter curve shape and falloff may be even more
> limiting. After you allow for non-synchronization (Nyquist limits..), you
> end up with typical system resolutions in the 40 to 50+ lpmm range for
> many consumer DSLRs, with lower often being better (as less noise) ;-)
So a theoretical 56 lp/mm would be cut to a lower value. Taking the old Leica
8 lp/mm printing standard, that would mean that many DSLRs would only give
their best prints at slightly over 4 times enlargement, and only with the
largest file size.
It occurs to me that simply photographing resolution charts would solve the
actual resolution puzzle. This would also be a better comparison with films.
So how come the only mention I have ever seen of photographing a resolution
chart with a digital camera has been done by Erwin Putts? He was testing the
Leica Digilux.
> As Carver Mead of Foveon noted, this is not likely to improve much, as we
> are already at the point where we are limited by light physics interacting
> with the chip, smaller means more noise and more light needed etc. So
> chips bigger than 16 MP are likely to be bigger than Foveon's trial 16 MP
> chip (which is 22x22mm square), not denser, and bigger will be better,
> meaning less noise, but also lower resolution per unit chip area than film
> This means a 64 MP sensor is likely to be Med Fmt ;p-)
My understanding was that the smaller Ym pixels were less light sensitive.
This caused more of a limit on setting in camera ISO values to avoid noise.
Going to a larger Ym pixel improves light sensitivity, but then file
resolution would be smaller for a given chip size, leading to the need for
larger chips. The other issue was adding a microlens layer to improve
sensitivity.
> there are a series of digital test charts which can be used to measure
> digital cameras directly, including an MTF analysis with free online
> software from the author (the chart, sad to say, is like $150 from the
> standards testing agency - unless you make a copy from the articles in
> mid-2002 in Brit. Jrnl of Photography showing one on lith film ;-) also
> distortion, color bias, and so on.
So where can I find published actual resolution numbers for digital cameras?
> Applied Optics V34 n4 pp. 746-51 of 1995 has Sitter et. al article on
> doing MTF and resolution analysis, including java source code by Jeff Kuhn
> at UTA at http://rsb.info.nih.gov/ij/plugins/mtf.html
>
> the above stuff could be used with standardized postscript created bar
> charts to actually measure MTF etc. on digital cameras, scanners, film
> scanners (lith film?), and so on.
I have an easier to use item in a printer control target. This is used for
press checks, and includes colour information for CMYK printing. If someone
might find one of these useful, I would be happy to upload it to an area of
my web site. It is an EPS file. Of course, this prints as a long strip of
information, and is mostly just useful for checking printed output.
I think an easier method is just to create some line pairs in Adobe
Illustrator. Print those out with a nice laser printer, then photograph the
results. Perhaps 20 lp/mm, 50 lp/mm, and 100 lp/mm would suffice.
> what I am saying is that it would be a minor effort for a digital magazine
> to setup a test suite, using standardized charts and test resources, and
> directly tell us how various cameras and scanners and so on performed.
>
> Ain't going to happen, though, first because it costs $$, and second
> because for every mfger it makes happy, it will make ten advertisers very
> very unhappy, esp. those who came in last. ;_) ;-) ;=)
Yeah, I don't think that resolution charts are in the best interest of the
direct digital camera manufacturers. I doubt that cost is an issue.
One last idea for you, and Rafe, and the rest. How about setting up a slide
duplication rig to hold a Kodak Q-60 target, and photograph that using a
direct digital SLR? Only need a bellows or macro lens then.
Ciao!
Gordon Moat
Alliance Graphique Studio
http://www.allgstudio.com
From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How big is the resolution on a normal 35 mm negative?
Date: 13 Dec 2003
Rafe B. rafe.bustin@verizon.net wrote
> This project will only go as far as the "volunteers" are
> willing to take it. If I can muster the time, I might call
> around to drum scanner vendors and see if they've
> got samples they're willing to submit. Aztek apparently
> doesn't want to play along.
>
> Again, my idea was to let folks cherry pick and send the
> very best they had to offer.
>
> Sending around a standard slide takes time and I don't
> really have an interest in making a huge project of it.
> I tried setting up something like that a couple years
> back on Mitch Leben's scan listserv and there was
> no interest.
>
> A shame that Tony Sleep lost interest in all that
> stuff -- his scanner site and listserv were very useful
> to me as I was getting started with film scanning.
> I don't see any current reviews there and I don't
> even know if the listserv is still running.
> rafe b.
> http://www.terrapinphoto.com
If you scroll down this page about halfway you will see a trio of
image crops comparing EOS 1Ds direct capture with drum-scanned and
CCD-scanned Provia 100F:
http://www.outbackphoto.com/reviews/equipment/Canon_1DS/Canon_1Ds_Diary_part1.html.
Same scene, same day, same lens, etc..
To my eyes, the digital shot is drastically superior to the film image
in this comparison, even when the film was drum scanned.
Brian
www.caldwellphotographic.com
From: Mxsmanic mxsmanic@hotmail.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How big is the resolution on a normal 35 mm negative?
Date: Sun, 14 Dec 2003
brian writes:
> To my eyes, the digital shot is drastically superior to the film image
> in this comparison, even when the film was drum scanned.
To my eyes, there's something wrong here. The drum scan is noisier than
the CCD scan, not cleaner; and both scans are worse than mine, and
blurry. The film shots aren't as well focused as the digital shot. The
1Ds was set to twice the ISO equivalent of the Provia; depth of field
was thus nearly twice as great with the digital shot, or camera shake
was half as great. Very strange.
From: Rafe B. rafe.bustin@verizon.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 6x7 med. format film scanner?
Date: Sat, 13 Dec 2003
"LightningBit.News" news@$NOSPAM*Lightningibit.com wrote:
>Bill,
>
>You always can lookout for a 2nd hand LeafScan 45.....
>they are old and big, but still beat modern scanners in quality
>
>they used to be (and still are) high end scanners (color or b&w;) for medium
>format
I don't know about "beat" but I'll agree that they were
and still are excellent machines.
The Leafscan 45 takes well over an hour to scan
a 35 mm image in color. It's limited to 2500 dpi for
medium format film, whereas the LS-8000 yields
4000 dpi. I've compared the two, in fact if you
check my "can comparison" site later today,
I may have comparative scans posted (Leaf 45
vs. LS-8000.)
Plus, the Leaf has a bizarre interface (IEEE-488)
and ancient drivers (eg. Win 3.1) due to its vintage.
If you can find a used one in good condition for
$1000 or under, it's probably worth going for.
rafe b.
http://www.terrapinphoto.com
scan snippets
http://www.terrapinphoto.com/jmdavis
From: rmonagha@engr.smu.edu (Bob Monaghan)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How big is the resolution on a normal 35 mm negative?
Date: 11 Dec 2003
very interesting, T.P., this is also what Pop Photo got in their digital
vs. film tests for 100 ISO film (brand unspecified), circa 40 MP (Pop
Photo march 2001 p.55) using color print film.
my suggestion is that you/we should invert the question, viz., given
kodak's ratings of its films, what scanner dpi is needed to effectively
use a given film speed or type?
i.e., 20-25 MP for 400-800 ISO speed color print film ~= 4000 dpi approx
40 MP for 100 ISO speed color print film ~= 6,000 dpi
150 MP for slow 25 ISO color slide film ~= 12,000 dpi scanner? ;-)
[i.e., 4x info, 2x density on each axis ;-)]
regards bobm
From: T P tp@nospam.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How big is the resolution on a normal 35 mm negative?
Date: Tue, 09 Dec 2003
"Sebastian Zamorski" info@sezam.dk wrote:
>I'm trying to find the benefits when it comes to a normal 35mm film and
>compare it to a digital picture.
>Could anybody tell me how big a resolution a 35mm negative has - I've heard
>something about 20 mill. pixels (?)
I think the most reliable way to make such a comparison is to look at
scans made with high resolution film scanners and decide what scanner
resolution represents the maximum practicable level of detail.
With a fine grained film such as Fuji Provia 100F, you will see a
significant difference when you go from 2800 to 3600 or 4000 ppi.
From 4000 to 5400 ppi the difference is not so great, and from 5400 to
6000 ppi shows little or no extra detail. So I suggest that the
maximum practicable level of detail is obtainable at 5400 to 6000 ppi.
At 6000 ppi, a 24 x 36mm negative would yield about 48 million pixels,
and at 5400 ppi about 40 million.
From: "William D. Tallman" wtallman@olypen.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Film scanners -- worth it?
Date: Wed, 31 Dec 2003
steven.sawyer@banet.net wrote:
> I have a Polaroid Sprint Scan 35 Plus. I bought it for around $200 off of eBay. I
> used to send my slides out for Cibachromes. This scanner has paid for itself many
> times over. Between it and my lowly Epson C82, I easily outperform the Frontier lab
> and nearly outperform Cibachromes. I cannot, however, rival C41 prints,
> but I haven't seen any consumer or low-end pro printing (i.e. Frontier)
> method beat C41 (prints from negatives) yet.
That's as in emulsion prints? Depends on the optics and use technique, like
anything else.
Try the Minolta DSE5400 and the Epson Stylus Photo 2200. Pretty good stuff.
Are you scanning in 48 bit color? The experts claim it cannot make a
visible difference, but I disagree. Does for me; YMMV.
Bill Tallman
From: Rafe B. rafe.bustin@verizon.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Film scanners -- worth it?
Date: Thu, 01 Jan 2004
"MikeWhy" mikewhy@my-deja.com wrote:
>"Mxsmanic" mxsmanic@hotmail.com wrote ...
>> trying to scan film with a non-film
>> scanner, throws 95% of the quality of the film image into the
>> wastebasket. Once you scan film directly with a film scanner, you'll be
>> amazed at just how nice those film images actually are.
>
>Not likely, but it's encouraging to think there's more detail hiding in my
>shoeboxes. By inference, then, film scanners are tossing 94% or more of the
>total information. They're good, but not that much better than my flatbed.
>
>In any case, your figures are without basis.
For kicks, have one of your better slides and/or
negatives scanned on a good 4000 dpi film scanner,
or maybe an Imacon or a drum scanner.
I'm on my third film scanner -- fourth if you include
the Epson 1640. As I write this, I am re-scanning some
3.5 year old negatives taken at Bryce Canyon. The
last set of scans -- on a Polaroid SprintScan Plus,
at 2700 dpi -- were pretty good, I thought. But this
newest round has *so* much better tonality, and
somewhat better detail as well.
Slide and film scanning is a tedious process, and
it pains me to think of all the work I've done on less-
than-optimal scanners. Think of what your time
is worth.
IMO, one should think of the scanner as no less
critical than the camera itself if image quality is
paramount. Matter of fact there's no camera in
my collection that costs anywhere near what that
LS-8000 set me back. It seemed like a lot of $
back then (and it was!) but I've never regretted
that purchase for a moment.
rafe b.
http://www.terrapinphoto.com
From: Rafe B. rafe.bustin@verizon.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Dynamic Range of films vs prints
Date: Sat, 25 Oct 2003
"Bart van der Wolf" bvdwolf@no.spam wrote:
>"Lourens Smak" smak@wanadoo.nl wrote...
>SNIP
>> Inversion isn't good enough; the (orange) mask layer has a varying
>> density too...
>
>The mask is designed to have a density+secondary absorptions that are a
>constant. Since secondary absorptions increase with the exposure time, the
>mask density reduces with exposure time.
>Subtract this constant and the mask and complementary secondary absorptions
>are gone. The best way of removing this constant mask color while scanning,
>is by increasing the green/blue exposure versus the red exposure.
Bingo.
A fellow named Dane Kosaka noted that NikonScan's
built in negative-inversion software was somewhat less than
optimal. He devised a nifty workaround. The simple variant
that I use goes like this:
1. Scan negatives as positives.
2. In the levels/histogram tool, click the
auto-levels button.
3. Now invert each color channel individually.
In step 2, you'll see that the red channel always
defines the high end of the tonal range, and has
the widest histogram, while the green and blue are
capturing a much narrower and generally lower range
of input codes. Setting the green exposure to +.6 units
and the blue exposure to +1.2 units is a good starting
point for normalizing the green and blue histograms.
In any case, following steps 2 and 3, one
generally ends up with a nicely inverted
image, with no clipping and the best possible
distribution of input values.
Dane's detailed notes on the treatment of
color negatives in NikonScan can be viewed at
http://www.marginalsoftware.com/LS8000Notes/LS8000Notes.htm
and
http://www.marginalsoftware.com/LS8000Notes/three_easy_ways.htm
rafe b.
http://www.terrapinphoto.com
From: Lourens Smak smak@wanadoo.nl
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Dynamic Range of films vs prints
Date: Sun, 26 Oct 2003
Rafe B. rafe.bustin@verizon.net wrote:
> I'm not sure how adding an extra step makes the process easier.
Well, you posted a "nifty workaround" in another message, for a problem
that you wouldn't have. ;-)
I actually tried this nifty workaround, and it is far from ideal.
Basically you scan a lot of the wrong information, which is tossed away
immediately after scanning...you then spread out the remaining bit of
data over the desired wider range. You end up with a reasonably
good-looking image, but with a gamut that is a lot smaller than desired.
(the histogram also clearly shows this...) A scan made that way will not
be suitable for very common post-processing tasks, with banding very
likely to happen for example; a bit of extra color-correction (or
cmyk-conversion...) for montage or layout purposes could also become
ugly quickly.
But for just a straight digital print the method would probably be
acceptable, if the workflow is kept 16-bit/channel for as long as
possible. But then, why not just print the negative?
Lourens
From: Robert Feinman robertdfeinman@netscape.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Dynamic Range of films vs prints
Date: Fri, 24 Oct 2003
The recent thread about transparencies vs negatives highlighted the
different perceptions that people have about this subject.
Lately I've done some scanning of color images with the new Minolta 5400
and have revised by opinion about the need for transparencies.
If you look at my tips section you will see several on capturing the
full range of densities and then scaling them to a suitable print.
The factor that seems to have been missed in the dynamic range
discussion is that it is not necessary to scale the densities uniformly.
As another of my tips shows it is also possible to adjust different
areas of the scene using different curves.
So, I've revised my opinion and I now think that color negative film is
probably easier to use and presents less restrictions when coupled with
a modern scanner having good Dmax and 16 bit processing.
--
Robert D Feinman
robertdfeinman@netscape.net
Landscapes, Cityscapes, Panoramas and Photoshop Tips
http://robertdfeinman.com
From: Rafe B. rafe.bustin@verizon.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Russell Williams on Slides v. Negs
Date: Fri, 24 Oct 2003
Here's a repost in its entirety, from a very clever fellow
named Russell Williams at Adobe. This post appeared
around four years ago on the now defunct Epson printer
mailing list run by Mitch Leben.
Those offended by any mention of "digital darkroom"
are advised to move on to the next post.
rafe b.
http://www.terrapinphoto.com
Date: Fri, 6 Aug 1999
To: epson-inkjet@leben.com
From: Russell Williams williams@Adobe.COM
Subject: Re: "Best" color slide film/color neg
I shot chromes for 25 years, and switched entirely to negatives
a couple of years ago. Others have noted the drawbacks of
machine prints. But I now make all my own prints on my EX and
that has completely changed my criteria. Pros and cons for me?.
1. At ISO 100 for enlargments to 11X14, I find the grain, sharpness,
color accuracy and saturation I can achieve comparable (Fuji Reala
vs. E100 or even Velvia). Differences: shadows are harder to scan
on slides without adding lots of noise. Low contrast areas on
negatives have a similar, but lesser problem. Skies often require
work to smooth them out in either case.
2. Contrast is much less of a problem with negative film and can
be easily added when necessary; it's much harder to reduce
contrast on a slide because areas of interest have often moved
into the toe of the film curve or gone completely black or white.
3. ISO 400 negative film produces excellent 8X10s and in
many cases very good 11X14s. ISO 400 slide film is terrible
by comparison. The only 8X10s with grain I found acceptable
from slides came from the new E200 pushed one stop.
4. Dynamic range is tremendously greater with negatives. I
never had much of a problem with slides' low exposure latitude --
I consistently got the exposures I wanted and covered tough
situations with occasional brackets.
The problem was that there was often no exposure that would
capture everything in a scene. One common problem: I spent
$70 on a good 2-stop graduated neutral density filter and spent
lots of time setting up tripod shots with it to avoid blown-out skies.
Once I switched to Epson prints, I tried just shooting a second
frame of just the sky and pasting it in with Photoshop. Major pain.
Negative film can usually capture the entire scene properly
in a single exposure (though it still requires "curves" work in
Photoshop) -- the sky or wall or whatever that looks completely
white in the lab print is actually there on the negative and can be
coaxed back with curves.
5. I have more trouble getting the colors right with negatives that I
scan myself because my scanner (like many others, apparently)
won't turn off its autoexposure when scanning negatives. That
means that my negative scans are inconsistent and prey to the
same problems (like subject failure) that plague really cheap
machine prints. For example, getting fleshtones right for portraits
of people wearing red shirts is a real pain -- the autoexposure
turns their skin cyan and I have to screw around with scan
controls and Photoshop curves. And any time you have to make
big color shifts you start getting second and third order problems
as you correct the first order ones. Others with more experience
in scanning than I also tell me that part of this problem is because
cheap scanners don't do a great job of getting rid of the orange
mask without introducing color shifts and crossover. PhotoCD
scans of negatives don't have this problem nearly as bad
(Kodak understands negative film pretty well).
SOME USEFUL INFO ABOUT SLIDE AND NEGATIVE FILM
A lot of the contrast / latitude / shadow noise / posterization issues
in scanning slide vs negative film come from these simple facts:
Slides capture a narrower brightness range of the original scene
and record it into a larger brightness range on the film. I don't
recall the exact numbers (and it varies by film), but slides capture,
say, a 5-stop range from darkest to lightest area of the original
scene, and translate that into a D of 3.6-4 on the film. Negatives
can capture a 7-8 stop range of the original scene and represent
that as a D range of 2.8 or so on film. There is more information
on the negative, but it is compressed into a narrower range. An
Epson print has a narrower range yet, so you have two mapping
problems: translate the D range of the film into the 24 or 48 bit
representation in the computer, then map that range onto the
paper.
If you take 8 bits per channel from a negative and then try to
significantly spread out some of it to make, say, highlight
details visible, you can get posterization.
So for slides, a scanner with a high D range is more important --
3.4-4. For negatives, it's more important to have greater bit
depth (e.g. 12 bits per channel) and pass that all to Photoshop
so you can manipulate it.
Russell
From: Rafe B. rafe.bustin@verizon.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Educating Stacey
Date: Thu, 23 Oct 2003
Stacey fotocord@yahoo.com wrote:
>And if you make a print that uses ALL of the info on print film it would be
>as flat as a pancake. A good print is going to have more contrast than the
>films contrast and as such only uses part of the range print film records
>anyway so what "advantage" does it have unless your interested in making
>very low contrast prints or can't be bothered to expose the film correctly?
No, you've got it completely ass-backwards, and even
so, you take another cheap shot at my skills.
I use every bit of the detail and tonal range that's on my
negatives. And every bit of dynamic range that my prints
can hold. To do otherwise would violate the few religious
principles that I hold.
As they say on NPR, "let's do the numbers" :
Prints have a Dmax of around 2.
Negatives have a Dmax of around 3.
Chromes have a Dmax of around 4.
Very rough numbers, I admit. Each of these
implies good exposures, ie., tonality that pushes
the medium to its limits.
Going from a chrome to a print implies a compression
of roughly 100:1. Going from a negative to a print
implies a compression of (merely) 10:1.
Ergo, the output tonal range of the print is in *no way*
constrained or limited by the tonal range of the original,
regardless of whether the original is a negative or a
chrome.
Ergo: prints from negatives can be just as snappy or
just as dull, as prints from slides.
Now here's the point you missed. Notice that while
negatives have lower Dmax than chromes, they also
have a much wider lattitude. So there's a lot more
compression that goes on in making the negative in
the first place.
Again, very rough numbers:
Negs: 7 stops (128:1) compressed into a Dmax of 3
compression = 128/1000 = 0.128
Chromes: 5 stops (32:1) compressed into a Dmax of 4
compression = 32/10000 = .0032
Ergo: Negs are 40 times more compressed than
chromes.
This is what makes chromes better for scanning
and printing -- IF -- you can keep from blowing them
out in the first place. And that's a very big IF, if
you're a landscape photographer shooting by
available light. (Oh yes, and if you're scanning,
you have to worry about the Dmax of the chrome
bumping into the scanner's limits, which is almost
never an issue with negatives.)
Me, I choose not to take that chance.
You clearly possess the greater skill, sufficient
to beat the laws of physics and optical science.
I stand humbled and in awe.
rafe b.
http://www.terrapinphoto.com
From: "Neil Gould" neil@myplaceofwork.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: future of 120/220?
Date: Tue, 13 Jan 2004
...
> I've been scanning film for six or seven years now and
> have never seen a moire artifact. Nor in half years' time
> using a 10D, or 18 months with a Canon G2.
I can only suppose that your not running into these problems has to do
with your subject matter. As much of the material that I get paid to shoot
has symmetrical patterns and such, I spend a good bit of time dealing with
these artifacts from digital images. It can be downright nerve wracking,
not to mention costly.
That said, one can minimize these artifacts when scanning film by
repositioning the target (i.e. rotating), changing scanning resolutions,
etc. Those options are not practical with a digicam.
Neil
From: "David J. Littleboy" davidjl@gol.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Mounted 6x7 vs. Unmounted (LS-8000)
Date: Thu, 12 Feb 2004
"Lunaray" yar@easystreet.com wrote
> Hello,
>
> I just shot my first roll of 6x7 transparencies and I'm wondering if I
> should have them mounted, or cut into strips? When I ordered my Nikon
> scanner, I ordered a mounted holder for it, so I can do either; any
> opinions?
How flat do slide mounts hold the film? If the answer is "not very", you'll
be very unhappy. The 8000 really wants the film held flat.
Mounts usually obstruct some of the image. And you won't be able to use the
glass holder (if you run into nastily warped slides). Unless you are going
to be projecting your slides, I don't see any reason to mount them. Many
people find that the regular strip holder (it's a bit of a rube Goldberg
that clamps and stretches the film) works fine; I'd think it would be better
than slide mounts for holding the film flat.
Just my opinion...
David J. Littleboy
Tokyo, Japan
From: "Gear>id O Laoi/Garry Lee" nospam_glee@iol.ie
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Coolscan 8000, more questions.
Date: Sat, 10 Jan 2004
I discovered how to deal with the film flatness problem in the 120 holder.
The film edge sits on thin rubberised strips which lose their natural
stickiness from time to time.
Just rub wet soap along these strips, then wash off with water, dry your
holder and off you go again. This is not in any manual but it works very
well.
The holder involves placing the film on these tracks and then closing and
putting a lateral pressure on it to flatten it. When these strips get a bit
dirty (every couple of months), just clean them. Othewise the film will not
be kept flat.
This scanner is superb in my opinion.
From: "-" xvvvz@hotmail.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Epson 4870 Press Release
Date: Tue, 06 Jan 2004
As first noted by a post from Erik de Goederen on photo.net, this was posted
on the MACminute website. I checked the Epson USA site and they didn't have
any press releases for the scanner and the website doesn't list the scanner
under the current product lineup.
Doug
---
Doug's "MF Film Holder" for batch scanning "strips" of 120/220 medium format
film:
http://home.earthlink.net/~dougfisher/holder/mfholderintro.html
Epson introduces two new scanners
January 6, 2004 - 00:32 EST Epson has announced two new scanners: the US$449
Perfection 4870 Photo and the $599 Perfection 4870 Pro. The former is
targeted to graphic arts and advanced amateur photographers; the latter is
built for advanced amateur photographers. Both offer flatbed single pass
operation; 4800 dpi optical resolution; 4800 x 9600 dpi maximum hardware
resolution with the Epson Micro Step Drive; 12800 dpi maximum interpolated
resolution; 48-bit color scanning with 3.8 dynamic range for transparencies;
16-bit internal/external; and 16-bit grayscale scanning. Both include
LaserSoft Imaging SilverFast 6.0 SE, Adobe Photoshop Elements 2, ABBYY
FineReader Sprint OCR and Epson Easy Photo Fix software. However, the pro
model also comes with Monaco EZColor and ArcSoft Suite applications. Both
scanners offer USB 2.0 and FireWire connectivity and are compatible with Mac
OS X 10.2 and higher.
From: "Bart van der Wolf" bvdwolf@no.spam
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Coolscan 8000ed (glass holder?)
Date: Fri, 21 Nov 2003
"David J. Littleboy" davidjl@gol.com wrote
SNIP
> The glass carrier can result in Newton's rings from the lower
> (non-anti-Newton) glass. It comes with masks that hold the film off the
> glass, but they only work for one frame at a time. I've cut my own mask that
> has enough extra width that I can get three 645 frames in at the same time
> for the various frame spacings on the various camera/120/220 combinations I
> use.
Just wondering, has anyone ever tried to quantify the loss of contrast due
to the addition of 4 (uncoated) air/glass surfaces? Each transition will
lose some 4-5% of the incident light, which has to go somewhere. The
contrast will be impacted most at the high film densities, they probably
have a higher RGB value than without glass, compared to low density film
areas.
Bart
From: kwinkler@sennheiserusa.com (Karl Winkler)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Medium Format Scanner which one?
Date: 5 Jan 2004
"Ray Paseur" ray@earplugsnon-aol.com wrote
> My experience with the Epson 2450 has been excellent. You'll probably have
> to fiddle around a lot with screen/print resolution issues, but the scanning
> part is fairly easy. One word of caution - the 2450 is not a production
> tool - at any reasonably high density it is very slow - several minutes per
> scan.
>
> Ray
I've used the 2450 for B&W; negatives (6x6cm) and the results are quite
acceptable. Far better than with color film! It is a bit slow, but as
someone else mentioned; so are the dedicated film scanners! The
Vuescan software is helpful, but the bundled version does not allow
for 16-bit scanning (or 48-bit in color). For that, use the standard
Epson software.
Here's a tip for B&W; scanning I picked up a few months back (a year
ago?) in Shutterbug:
Scan the image as a B&W; positive, using the 16-bit setting, and do not
pre-set any levels or curves.
Once you have the image scanned that way, it should look like a flat
(low-contrast) negative on the screen. Using Photoshop, do the first
pass of a LEVELS adjustment (Image>Adjust>Levels), by moving the right
(white point) slider to the left until it just touches the first real
part of the image in the histogram. Hit OK.
Step 2: invert the image (turn it from a negative to a positive) by
using Image>Adjust>Invert. Now your image should look like a
washed-out positive.
Step 3. Do another LEVELS adjustment, again, moving the right-hand
slider (white point) so that it just touches the start of the image in
the histogram.
Step 4: before hitting OK, also adjust the mid-point slider around so
that the overall image takes on the tonal balance you are expecting.
Then hit OK.
Step 5 (optional) go to Image>Adjust>Curves and set up a curve to more
finely adjust the tonal balance of your image. Hit OK.
Step 6: Go to Image>Mode>8-bit to now convert the 16-bit image to
8-bit. Now (in Photoshop 7 or earlier) you can do layers or whatever
other local tonal control.
What this does is much more accurately preserve the micro-tonal levels
in the image, allowing for much smoother gradients and shadow detail.
If you had done these drastic adjustments with an originally 8-bit
image, you could have lost a lot of information, causing gaps in the
histogram.
Try it!
-Karl Winkler
http://pages.cthome.net/karlwinkler
From: "-" xvvvz@hotmail.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Medium Format Scanner which one?
Date: Mon, 05 Jan 2004
> Well, so are the "production tools". ;-)
If you have a strip adapter, the 2450 is much more productive. You can then
walk away and let it batch scan multiple negatives while you do other
things. Use some scrap mounting board that is about 1mm thick (which you
probably have sitting around the darkroom already!). I have a mini tutorial
on how to setup your software to batch scan on my website.
Doug
--
Doug's "MF Film Holder" for batch scanning "strips" of 120/220 medium format film:
http://home.earthlink.net/~dougfisher/holder/mfholderintro.html
From: J C null@nowhere.net
Newsgroups: rec.photo.misc
Subject: Re: Sufficient scan resolution
Date: Thu, 01 Jan 2004
That might be your contention, but it is simply not true.
Here's another example, How do you think CMYK 300 dpi digital image is
printed on a printing press?
Well, the printing press has to have 4 plates for the image, one plate
for each ink color. Each plate contains a halftone of the color. Each
plate can have a resolution of 150 haltone lines per inch. Now take a
look at a color photograph printed in a magazine. The halftone dots
containing color are laid down in a rosette pattern (this is so that
the ink colors are offset from each other, because if they were
directly on top of each other the image would be muddy).
So, now if each of the 4 CMYK colors can print with a 150 line screen,
then I'd suppose that 4 x 150 = 600 discrete image points per inch in
a printed color image. And by comparison, if you used a 75 line screen
for color (so that 4 x 75 = 300), then you'd get a noticeably lower
quality image. The difference in quality between a color image printed
in a newspaper and one printed in an artbook are extreme examples. The
news paper uses about a 65 line screen and crappy paper and an artbook
a 150 or higher screen and excellent paper.
There's even more to it than this, but....
Your "300 dpi is what the human eye can resolve" argument is WRONG.
Your sticking to your argument is simply limiting you frorm
understanding the process.
-- JC
From: "-" xvvvz@hotmail.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 6x7 med. format film scanner?
Date: Tue, 09 Dec 2003
Many people find the Epson 2450, 3170 and 3200 scanners meet their needs for
medium format. Since you are doing the larger 6x7 size MF film, these
scanners are worth a look. The 3200 is very inexpensive now at $242ish from
Amazon with free shipping and after rebate. The upcoming 4870 will be nice
but it may be a few months before it becomes available in stores and rumor
has it the 4870 will be between $400-$500.
Doug
--
Doug's "MF Film Holder" for batch scanning "strips" of 120/220 medium format film:
http://home.earthlink.net/~dougfisher/holder/mfholderintro.html
From: fotoralf@gmx.de (Ralf R. Radermacher)
Newsgroups: rec.photo.equipment.medium-format
Subject: great improvement for scanning MF with Epson 2450
Date: Sat, 31 Jan 2004
Everybody who has ever scanned film with a 2450 will know the problem.
The film holders are a bad joke and the film, heated and dried out by
the heat from the transparency lightsource, winds itself out of the
holder. This is particularly bad with the MF/4x5 holder where the scan
times are long enough that the film can move around a lot during the
eternity it takes to complete a single scan.
Now, I've just got myself the equivalent holder from the 3200 which is
fitted with a frame to securely clamp the film strip down in order to
hold it flat and prevent it from moving during the scanning process.
What a difference!!!
Bought it from one of the two official suppliers of Epson spare parts,
here in Germany, and paid a mere 11 euros, P&P; included!
Sometimes one is really left wondering why one hasn't thought of such
simple solutions much earlier.... :-)
Ralf
--
Ralf R. Radermacher
From: "Roger N. Clark (change username to rnclark)" username@qwest.net
Newsgroups: rec.photo.equipment.large-format,rec.photo.digital,rec.photo.technique.nature,comp.periphs.scanners
Subject: Re: scanning large format: to the limit (and beyond)
Date: Sun, 30 May 2004
Tom Monego wrote:
> Trying to find the name, he wrote a couple of articles in Yahoo's Epson Wide
> Format group stating that he was printing at 720ppi and 1440ppi. Scans 4x5's to
> 4000ppi (Aztek drum) and prints at 16x20 monochrome. He teaches at the New
> School in New York and tours with Cone Editions doing workshops. Been going
> through archives etc. trying to get the name. But he is questioning the status
> quo and backing up his theories with work. I find his work interesting and got
> me looking at prints at 240, 300, 360, and 720ppi. I personally saw a
> difference at 240 to 360ppi, 720 just didn't seem that much different and the
> file size got huge.But I'm working at 360ppi more now. Oh yes using an Epson
> 7600 with the pieziochrome inksets.
Tom,
Ive done experiments with different ppi prints, going up to
600ppi. see:
http://www.clarkvision.com/imagedetail/printer-ppi
I've done blind tests of people sorting prints in order
of sharpness. In good "office" or outdoor lighting almost
everyone gets the sequence correct, choosing the 600 ppi
print as sharpest. As light levels drop, it becomes more
difficult. I need to present the statistics on the
above page.
Roger
From: Hank Graber [hgraber@narrativerooms.com]
Sent: Sat 5/22/2004
To: hasselblad@kelvin.net
Subject: Re: [HUG] A Modest Discussion
If you scan a resolution target (like the USAF Resolution Test Pattern)
on a prosumer scanner you will find the real optical resolution to be
something less then 2500 dpi. The best drum scanners cannot resolve
more then 8000 dpi.
In any case comparing pixels to scanner dpi doesn't tell you anything.
You really need to compare the final output that will be used to
measure the suitability of any system analogue or digital. If your
final output is a 35mm slide for projection then none of the DSLR's
will come close to matching the quality of your 35mm transparency. If
your final product is a CMYK 8 1/2" by 11" 133 line screen web offset
printed page, well then the Canon 1DS is a suitable replacement for MF
film for a lot of applications with no sacrifice in quality and a more
flexible workflow.
Regards,
Hank Graber
Q.G. de Bakker wrote:
> Stein wrote:
>>> I used a pencil and a chip wrapper to calculate the equvalent for my A
> 12 backs* - using a 56 mm x 56 mm final size of negative. I came out with
> something like 30 Megal pixels, based on the ratio of capture areas. But was
> I correct? Is a colour negative film emulsion going to capture the same
> number of small visible elements, call them what you will, that is expressed
> by the silicon surface of the electronic device? Is there a direct comparison?
>
> Well, scanning 56x56 using one of the barely affordable scanners produces in
> excess of 80 MP, all pixels being good, no "empty magnification".
> Increasing scan resolution you will eventually cross the limit between
> scanning information captured on film (i.e. the useful bit) and scanning
> details of the grain/dye cloud.
> Seeing what a grain of silver or cloud of dye actually looks like up close
> could be interesting, but i doubt the followers of the Digido would allow
> that. But the 80+ MP stands firm. At 4000 dpi scan resolution we haven't
> crossed the limit yet.
From: Raphael Bustin rafe.bustin@verizon.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: real photography only 30 lpmm?
Date: Tue, 29 Jun 2004
Robert Feinman robertdfeinman@netscape.net wrote:
>Anyway, if you look at my results with the Minolta 5400 scanner I was
>able to get 12-18x enlargements which look much better than would have
>been done with a wet darkroom.
>I think it may be a combination of new lens design, improved film and
>much better detail capture when using a scanner. So I revise my
>estimate upwards to about 12x when everything works optimally.
Yep, this is what a lot of us have seen.
Film scanning and digital darkroom
making sharper prints, and with better
tonality, than a wet darkroom.
>When using my Pentax 67 and the Epson 4870 scanner the overall results
>don't appear any better than what I got with 35mm. Perhaps older lens
>design, more camera vibration (even with mirror up) and a much lower
>resolution scanner (in spite of advertised numbers).
>So depending on the workflow, medium format may not be noticeably better
>than 35mm.
>You can see the samples in the tips section of my web site.
I suggest it's the 4870 that's the weak
link here. Have you had one of your
good MF negatives scanned on a top-
notch film scanner? I think that might
cause you to change your tune.
I see noticeable differences between
35 and 645. But that's because I scan
both of them on the same 4000 dpi
scanner.
How does your scanner hold up against
comparable samples at my site below?
Back when I was scanning MF on an
Epson 1640 (supposedly 1600 dpi, but
seemingly half of that) I was somewhat
disappointed with MF.
rafe b.
http://www.terrapinphoto.com
scan samples site:
http://www.terrapinphoto.com/jmdavis/
From: Raphael Bustin rafe.bustin@verizon.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Mental rigor. Re: MF velvia > 300 MP? ;-) Re: why wet prints > 300 dpi
Date: Wed, 23 Jun 2004
Fil Ament 2Beseen@Mixelpix.twixt wrote:
> Yes well I would expect that to be the sole thread and attribute
>you would gasp at. How about grain structure? Which was my
>premise. The two films do have lots of saturation, they also scan&
>publish well ;-) I know from experience. As for objective reality
>thats for discussion, not perhaps for the artwork.
In general, chromes, and in particular, contrasty chromes,
are more difficult to scan than any color negative film.
Or let's just say that it takes a good scanner and a
skilled operator to capture the 3.5+ decades of contrast
on a good slide. If you're using color management you'll
want to use a particularly large color space, at the very
least AdobeRGB.
So I wouldn't claim that slides "scan well" at least not
from personal experience, compared to the obvious
alternative of print films.
Grain structure of Velvia? None that I can discern, but
then I'm not Dave L. I cut my teeth on Tri-X and Accufine.
When I look at digicam pix or scans from chromes, I
have to suspend my disbelief in the lack of grain... it
just doesn't look right {G}. OTOH, Reala suits me fine.
rafe b.
http://www.terrapinphoto.com
From: Hank Graber [hgraber@narrativerooms.com]
Sent: Sat 5/22/2004
To: hasselblad@kelvin.net
Subject: Re: [HUG] A Modest Discussion
If you scan a resolution target (like the USAF Resolution Test Pattern)
on a prosumer scanner you will find the real optical resolution to be
something less then 2500 dpi. The best drum scanners cannot resolve
more then 8000 dpi.
In any case comparing pixels to scanner dpi doesn't tell you anything.
You really need to compare the final output that will be used to
measure the suitability of any system analogue or digital. If your
final output is a 35mm slide for projection then none of the DSLR's
will come close to matching the quality of your 35mm transparency. If
your final product is a CMYK 8 1/2" by 11" 133 line screen web offset
printed page, well then the Canon 1DS is a suitable replacement for MF
film for a lot of applications with no sacrifice in quality and a more
flexible workflow.
Regards,
Hank Graber
Q.G. de Bakker wrote:
> Stein wrote:
>
> I used a pencil and a chip wrapper to calculate the equvalent for my A
> 12 backs* - using a 56 mm x 56 mm final size of negative. I came out with
> something like 30 Megal pixels, based on the ratio of capture areas. But was
> I correct? Is a colour negative film emulsion going to capture the same
> number of small visible elements, call them what you will, that is expressed
> by the silicon surface of the electronic device? Is there a direct
> comparison?
>
> Well, scanning 56x56 using one of the barely affordable scanners produces in
> excess of 80 MP, all pixels being good, no "empty magnification".
> Increasing scan resolution you will eventually cross the limit between
> scanning information captured on film (i.e. the useful bit) and scanning
> details of the grain/dye cloud.
> Seeing what a grain of silver or cloud of dye actually looks like up close
> could be interesting, but i doubt the followers of the Digido would allow
> that. But the 80+ MP stands firm. At 4000 dpi scan resolution we haven't
> crossed the limit yet.
From: "Roger N. Clark (change username to rnclark)" username@qwest.net
Newsgroups: rec.photo.equipment.large-format,rec.photo.digital,rec.photo.technique.nature,comp.periphs.scanners
Subject: scanning large format: to the limit (and beyond)
Date: Sun, 23 May 2004
I thought I would pass on my experience so far scanning large format
images with my new Epson 4870 scanner. Later, I'll post some comparison
images so you can see how good this scanner is. So far my experience
is that it is close enough to drum scans I've had done that I will
use it for almost all my work. It is not quite as sharp on Fujichrome
Velvia as the drum scans I've done, but careful sharpening overcomes
the limit for the most part. However, because it can do 16-bit
compared to the drum scans, I believe it is better because I can
recover more shadow and highlight details.
The problem is that at 16 bits/channel, no scanner software I've used
can scan the full width of a 4x5 transparency at anywhere near the rated ppi
of the scanner. My requirements are 3200 ppi minimum (the scanner does
4800 ppi). In my testing, 3200 ppi gets information important to my
images that 2400 ppi loses. I've tested Epson scan, Silverfast, and
Vuescan, all the latest versions. The limit, as discussed in a recent
rec.photo.large-format and comp.sys.scanners thread, is due to
a firmware limitation in the scanner limiting total bytes per line, and
that limit does not allow the full width of 4x5 to be scanned at 16 bit
except at 2400 ppi. Vuescan, for example, reads 4800 ppi
if you request 3200 then downsamples. Epson scan was the one I used.
I can do 3200 ppi, 16-bits/channel and a 3.4-inch line width.
With ICE turned on, such a scan takes about 1 hour. Then I move the
box over the the other half of the image, with lots of overlap, and
scan a second time, another hour. I make sure the settings are exactly
the same for the two scans. The resulting images are
about 14,820 x 11,740 pixels and 1 GByte. This joining procedure
went well in photoshop CS on a 1.8 GHz win XP box with 2 GB ram
and 600 GB disk.
I combine the two halves in photoshop. The intensity levels match
essentially perfectly: you can't see the join line at all. I have
found that some, but not all scans mis-register by about a half pixel
in blocks of a few hundred scan lines, meaning one block will be dead
on, the next off slightly. I erase the edge of the overlap
image to so it is not straight, add some feathering, and follow
darker portions of the image if possible and the the images go together
without a possibility of finding a join line.
Then, to really push the limits, I mosaiced two such 4x5 images
into a panorama. The result: 23,380 x 11,820 pixels. But here is
where it really became difficult. The combined file size in
photoshop, keeping them as layers was over 2 GBytes and photoshop
would not save the file when I tried. Fortunately it did not
crash either, so I had to feather the join line and merge the
images before the file size dropped below 2 GBytes. I'm using NTFS
(file system) so files can be larger than 2 GB, but photoshop
would do it in standard photoshop format. If someone knows of a way
for photoshop to save such a file (and read it in again later),
please let me know. The final image is 1.62 GBytes.
The problem with mosaicing images is the lens distortions. If someone
knows of software that will do the mosaicing on such large images,
please let me know.
Anyway, Thanks for those in previous threads who helped me work out
the details and limits of the scanner. I will just have to sigh, and
scan in pieces! So here is the final image:
http://www.clarkvision.com/galleries/gallery.NEW/web/colorado.fall.c10.01.2003.L4.9536.a+b.c.791.html
If people want more information on the procedure, I'll be happy to
provide it.
Roger Clark
http://www.clarkvision.com
From: Hill, Michael D. CIV (FASO) [michael.d.hill@navy.mil]
Sent: Fri 4/23/2004
To: panorama-l@sci.monash.edu.au
Subject: scanning long film
>Yes, you can scan the 175 neg in one pass. I use an Epson 3200, with a
homemade mask/neg holder. It works quite well, although the software limits
the size of the scan to under 30000 dots on any axis, so you can really only
scan for a 30" print at a max of 999 dpi. Bummer...
I just got the MF film holder http://home.earthlink.net/~dougfisher/holder/mfholderintro.html,
which allows a 202mm scan. Have not yet tried it with my 4870. Even if the Epson driver balks at that
length, I'm confident that Vuescan will scan it at max resolution.
Denis
From leica user group:
Date: Tue, 18 May 2004
From: "Jeffery Smith" jls@runbox.com
Subject: RE: [Leica] Scanning silver halide films - which one scans best? - OT
To: "'Leica Users Group'" lug@leica-users.org
It seems that not all scanners are born equal. I can perform the exact
same procedures using the same scanner with different films and one film
does very well while another scans very poorly. The Nikon Coolscan III
doesn't seem to like older fast films. The highlights block up unless
the film is underexposed, and then the mid-tones are grainville. Since
Nikon doesn't volunteer this information in their literature (I'm being
sarcastic...it came with no literature or instructions other than a card
telling you how to set it up hardware-wise), it is a game of hit and
miss.
Jeffery Smith
New Orleans, LA
End of Page