Scanning 8x10 glass negatives?
John Ackermann
are the most likely ones I can think of to have ideas on this topic.
My company, NCR Corporation, has a huge historical archive. Among other
things, we have what we believe is the world's largest collection of 8x10
glass negatives (over 100,000 of them). We also have 100,000 or so
hand-tinted glass magic-lantern slides.
We've recently entered into an agreement with the local historical society to
make this archive accessible to the public. One of our first objectives is to
preserve these negatives and slides by converting them to digital format.
That brings me to my questions.
First, we need to find a scanner that can handle 8 x 10 inch glass negatives.
I'm concerned that the thickness of the glass may make it difficult to use a
standard scanner. Any suggestions on a scanner that would be good for this
purpose?
Second, we're puzzling over the resolution and compression we should use for
scanning and storing these negatives. If I've done the math right, scanning 8
x 10 inches at 300dpi and 8 bits of grayscale (these are black and white)
results in a 72MB image. Times 100,000 images, that's a lot of CD-ROMs!!!
Since preservation is our prime goal, we don't want to lose noticeable detail
in the digitization process. Any suggestions on what scanning resolution, and
compression format, we should consider to reduce storage size while not losing
noticeable quality?
The same questions apply to the smaller (about 3x4 inch image
area) magic-lantern slides, which have the added challenges of being in color,
and of having the emulsion sandwiched between two pieces of glass, so the
image will be elevated a bit from the scanner bed.
Thanks for any insight anyone can offer.
John Ackermann
john.ac...@daytonOH.ncr.com
Ed Hamrick
>First, we need to find a scanner that can handle 8 x 10 inch glass
negatives.
>I'm concerned that the thickness of the glass may make it difficult to use
a
>standard scanner. Any suggestions on a scanner that would be good for this
>purpose?
You might try a flat-bed scanner with a transparency adapter, like the
HP 6100C. It'll cost you a bit, but there's a good chance it will work.
This scanner is somewhat unique in it's ability to focus on objects that
are a bit displaced from the scanner bed. As an example, it will scan
coins, pens, and even hands quite well. I suspect that this relatively
large depth of field will work well with your application.
Since you're planning to scan 200,000 slides, the cost of a scanner
or two is trivial compared to the labor costs of doing this. You
might as well buy a few different scanners to find out which one
works best. Scan speed is no doubt also important when scanning
200,000 slides, so measuring the scan speed of different models
is advisable. A 10% difference in scan speed may mean a 10%
difference in scanning labor.
>Second, we're puzzling over the resolution and compression we should use
for
>scanning and storing these negatives. If I've done the math right,
scanning 8
>x 10 inches at 300dpi and 8 bits of grayscale (these are black and white)
>results in a 72MB image. Times 100,000 images, that's a lot of CD-ROMs!!!
It's doubtful that the images have 300 dpi of resolution in their emulsion
if they're fairly old. In addition, you may want to preserve the color tint
of the negatives, especially if 100,000 of them are hand-tinted. Saving the
color information of tinted negatives doesn't add much to JPEG file size
since the tint is relatively constant over the negative.
I recommend you do testing to determine the minimum resolution that you
can use, and do extensive testing to determine the minimum JPEG compression
you can use. These two parameters will determine the file sizes. 8x10
inches at 200 dpi is about 10 MBytes and a JPEG file of this should be
about 1 MByte. The degree of JPEG compression is very image dependant.
>Since preservation is our prime goal, we don't want to lose noticeable
detail
>in the digitization process.
"Noticeable detail" means different things to different people. A purist
might want to scan at 2400 dpi and save the images in 24-bit TIFF, which
would result in 1.3 GBytes per uncompressed TIFF file.
As a practical matter, I suspect 100 dpi with JPEG compression will
provide more detail than anyone could see with a magnifying glass anyway.
If the slides are old, it's possible that there's only 100 dpi of
useful information before you're just scanning the film grain.
At 250 KBytes per slide, you might fit 20,000 slides per CD and get
the whole collection of 200,000 slides on 10 CD's.
Organizing the collection so it is accessible in a useful way will
also be quite a challenge.
Regards,
Ed Hamrick
Stan Kazmiruk
quality" dual-mode flatbed scanner. It scans reflective materials on top in
the usual way and scans transparent materials in a pull-out drawer below the
scan head. It comes with holders for a variety of film sizes up to 8"x10".
The scanner has excellent dynamic range (3.4) and produces much cleaner
shadow areas (lower noise) than other models I've worked with.
So far I haven't tried it with glass plates (I think I have a few stored
away somewhere . . .). When using the film drawer, the ScanMaker 5 scans
from the emulsion side of the film or transparency with no intervening
glass, but I'm not sure whether the added thickness of the glass plate would
affect the focus. If the thickness of the glass were an issue, it should be
possible to modify the film holder to lower the glass plate into the optimum
position. Unlike my 35mm film scanner, I don't think the ScanMaker 5 has an
autofocus mechanism, but perhaps its optics can accomodate some variation in
the position of the film emulsion. The manufacturer might be able to offer
some advice, but I think the only real way to find out is to try it.
The scanner is high quality but not very fast. I haven't done the math, but
I think if you wanted to scan 100,000 images, you would need several
scanners, workstations and operators working for a long time.
I think 300 dpi is a pretty good choice for your 8x10's. Fortunately, your
calculations are out by a factor of 10. An 8"x10", 8-bit B&W image at
300dpi will produce a 7.2 MB file. If you saved them as TIFF's, using
lossless LZW compression, the file sizes would drop to somewhere in the 3 to
6 MB range, depending on content. With JPEG, you can select the level of
compression. There is, of course, a trade-off between quality and
compression, but for most images, you can get, say, 10 to 1 compression with
barely noticeable changes to the image. From what I've seen, JPEG artifacts
are most noticeable in things like line artwork (eg. - pen and ink drawings)
and least noticeable in photographs such as landscapes.
The same comments would also apply to your 3"x4" color slides. For these
slides, 300 dpi scans would produce images of 900 x 1200 pixels. That's
comparable to the best consumer-grade digital cameras today, but it may not
capture all the detail in the original slide. If the originals are good
quality, you might want to use at least 600 dpi. For color images, that
would increase the file sizes to nearly 13 MB before compression.
It also occurs to me that, for your 3" x4" slides, it might be possible to
modify or fabricate a film holder for the ScanMaker that would allow
scanning 4 slides at a time.
Stan
John Ackermann wrote in message ...
>I realize this question isn't quite on topic for these groups, but you
folks
>are the most likely ones I can think of to have ideas on this topic.
>
>My company, NCR Corporation, has a huge historical archive. Among other
>things, we have what we believe is the world's largest collection of 8x10
>glass negatives (over 100,000 of them). We also have 100,000 or so
>hand-tinted glass magic-lantern slides.
>
>We've recently entered into an agreement with the local historical society
to
>make this archive accessible to the public. One of our first objectives is
to
>preserve these negatives and slides by converting them to digital format.
>That brings me to my questions.
>
>First, we need to find a scanner that can handle 8 x 10 inch glass
negatives.
>I'm concerned that the thickness of the glass may make it difficult to use
a
>standard scanner. Any suggestions on a scanner that would be good for this
>purpose?
>
>Second, we're puzzling over the resolution and compression we should use
for
>scanning and storing these negatives. If I've done the math right,
scanning 8
>x 10 inches at 300dpi and 8 bits of grayscale (these are black and white)
>results in a 72MB image. Times 100,000 images, that's a lot of CD-ROMs!!!
>
>Since preservation is our prime goal, we don't want to lose noticeable
detail
Don Stauffer
As long as you scan with the emulsion side down, against the scanner
glass, the glass of the neg will create no problem. You must, of
course, have a white backing, but I would imagine the standard cover
would have enough mechanical compliance to fit well against the back of
your neg.
Now, the magic lantern slides will be something again. The glass
thickness may create SOME defocus and hence loss of resolution. The
amount will depend on the numerical aperture of the scan system so there
is almost no way to tell except to try it.
As far as resolution and compression, 150 dpi does a pretty good job and
reduces file size by a factor of four. Try one each way and see how you
like it. If archiving is the objective, I suggest you use a TIFF
format. There IS some compression, and it is a lossless compression so
no info is lost. It is a recognized file format that is likely to be
around for a long time.
--
Don Stauffer in Minneapolis
home web site- http://home1.gte.net/stauffer/
home email- stau...@gte.net
work email- stau...@htc.honeywell.com
Bob Wheeler
John Ackermann wrote:
>
>
> Since preservation is our prime goal, we don't want to lose noticeable detail
> in the digitization process. Any suggestions on what scanning resolution, and
> compression format, we should consider to reduce storage size while not losing
> noticeable quality?
>
> The same questions apply to the smaller (about 3x4 inch image
> area) magic-lantern slides, which have the added challenges of being in color,
> and of having the emulsion sandwiched between two pieces of glass, so the
> image will be elevated a bit from the scanner bed.
>
>
Other posters have covered most of the salient points. I second the recommendation
that you look at the Microtek 5.
The scan density you choose depends on both the final use and the amount of detail
possible in the slides. Your choice of 300dpi is a good choice for printing, since
this will produce a "contact" print with as much detail as the human eye can see.
If you anticipate an enlarged print, then you should multiply the enlargement
ratio by 300.
If you think it possible that someone will want to transfer the images to film
for projection, then 300dpi is much too small. Even a modest projection of a 3x4
inch slide onto a 3x4 foot screen will require a 3600dpi scan -- the Microtek has
a maximum of 1000dpi. The scan you choose depends on the amount of detail in the
magic lantern slides. One way to determine this is to project one of the slides,
pick several small details that are marginally defined, and then scan the slide at
various scan rates until the zoomed on-screen details become marginal. I think it
unlikely that you will find detail requiring more than about 1000dpi -- this
corresponds to about a 20 lpm resolution on the slide, which is about what modern
8x10 films produce.
--
Bob Wheeler --- (Reply to: bwhe...@echip.com)
ECHIP, Inc.
peter
negatives.
> I'm concerned that the thickness of the glass may make it difficult to
use a
> standard scanner. Any suggestions on a scanner that would be good for
this
> purpose?
> Second, we're puzzling over the resolution and compression we should use
for
> scanning and storing these negatives. If I've done the math right,
scanning 8
> x 10 inches at 300dpi and 8 bits of grayscale (these are black and white)
> results in a 72MB image. Times 100,000 images, that's a lot of
CD-ROMs!!!
> Since preservation is our prime goal, we don't want to lose noticeable
detail
> in the digitization process. Any suggestions on what scanning
resolution, and
> compression format, we should consider to reduce storage size while not
losing
> noticeable quality?
> The same questions apply to the smaller (about 3x4 inch image
> area) magic-lantern slides, which have the added challenges of being in
color,
> and of having the emulsion sandwiched between two pieces of glass, so the
> image will be elevated a bit from the scanner bed.
Hi John,
I'm using the Agfa Duoscan for this kind of work. It can handle plates up
to 9 1/2 to
12 inches. The same sizes also for reflective pics.
The optical resolution should be 1000 dpi. So you need a computer with a
lot of memory!
RAM and also harddisk.
For one 8x10 with 1000 dpi appr. 60 MB.
But for the hand-painted lantern-slides you need not more than 300 dpi. As
RGB-file
its appr. 3MB.
There is no big problem with the sandwitched slides. The handpainted pics
dosn't have
a high resolution. So a little difference in the high doesn't make any
trouble. Look for the
side with the thinnest glas and frame. Put this side to the glass-plate
from the scanner.
You can always exchange the view with the imaging-program.
You have to safe the datas with an image-format without the loss of
resolution. For
example in TIFF.
But there is always one problem: You can see this negatives also after 100
years without
any tool. You cannot be shure, this works also with datas in the next 100
years.
Peter
Andrzej Irla
associate digital photography with unnatural alteration
of images, it is becoming a great tool to simulate the
dark-room process of enlarging and chemical printing, and
it has significant application in large format photography.
For example, EverColor prints, based on digital technology,
are superb quality and long lasting. So the question is not
"if" but "when" digital technology becomes a very popular
way of printing?
If your goal is preservation, you need to produce a digital
master copy of each image. For the purpose of preserving every
possible detail from your physical medium, the resolution of
the digital master copy file should be matched to the maximum
resolution of your glass negatives. For example, assuming that
your glass negatives have maximum resolution of 20 lpm
(lines per millimeter) then you need to scan that image at
(20 x 2 x 25.4 = 1016) dpi (dots per inch) to achieve
one-to-one correspondence between digital and physical dots.
A master copy file with 1000 dpi will give you ((8 x 1000) x
(10 x 1000) = 80,000,000 bytes) or 76.29 MB file of gray scale
image. Considering CD Rom capacity of about 650 MB, you could
store only about 8 image files per CD (TIFF format). I realize,
this would take a very large amount of resources, but I don't
believe that each of the 200,000 negatives is a masterpiece
worth such effort. I would first evaluate all images and
select those with any artistic values, and scan them with the
highest resolution. The rest that have some historical value
could be scanned with lower resolution.
The only file format that you should store master copies are
the ones that use lossless compression algorithm for example
TIFF, PNG or BMP. Do not use JPEG format because it uses a
lossy compression algorithm, and, depending on level of
compression, it will corrupt some of the pixels permanently.
You may be tempted to buy one of the cheapest general purpose
flatbed scanners available on the market. But, keep in mind
that scanners are like lenses. The better the scanner, the
higher price you pay. For photographic art applications, avoid
general purpose scanners that are mainly designed for scanning
prints or vacation snap-shoots. Support companies like Agfa,
Umax, or Microtek who recognized our photographic needs and
spent money on research to release flatbed models that can
satisfy the needs of large format photographer. Especially,
the Mikrotek SnacMaker 5 seems to be a wonderful piece of
equipment.
Regards
Andre Irla
John Ackermann wrote:
>
> I realize this question isn't quite on topic for these groups, but you folks
> are the most likely ones I can think of to have ideas on this topic.
>
> My company, NCR Corporation, has a huge historical archive. Among other
> things, we have what we believe is the world's largest collection of 8x10
> glass negatives (over 100,000 of them). We also have 100,000 or so
> hand-tinted glass magic-lantern slides.
>
> We've recently entered into an agreement with the local historical society to
> make this archive accessible to the public. One of our first objectives is to
> preserve these negatives and slides by converting them to digital format.
> That brings me to my questions.
>
> First, we need to find a scanner that can handle 8 x 10 inch glass negatives.
> I'm concerned that the thickness of the glass may make it difficult to use a
> standard scanner. Any suggestions on a scanner that would be good for this
> purpose?
>
> Second, we're puzzling over the resolution and compression we should use for
> scanning and storing these negatives. If I've done the math right, scanning 8
> x 10 inches at 300dpi and 8 bits of grayscale (these are black and white)
> results in a 72MB image. Times 100,000 images, that's a lot of CD-ROMs!!!
>
> Since preservation is our prime goal, we don't want to lose noticeable detail
> in the digitization process. Any suggestions on what scanning resolution, and
> compression format, we should consider to reduce storage size while not losing
> noticeable quality?
>
> The same questions apply to the smaller (about 3x4 inch image
> area) magic-lantern slides, which have the added challenges of being in color,
> and of having the emulsion sandwiched between two pieces of glass, so the
> image will be elevated a bit from the scanner bed.
>
Michael Greer
Stan Kazmiruk wrote:
> The Microtek ScanMaker 5 is a recent model, 36-bit, high-end "professional
> quality" dual-mode flatbed scanner. It scans reflective materials on top in
> the usual way and scans transparent materials in a pull-out drawer below the
> scan head. It comes with holders for a variety of film sizes up to 8"x10".
> The scanner has excellent dynamic range (3.4) and produces much cleaner
> shadow areas (lower noise) than other models I've worked with.
The Scanmaker 4 "seems" to be very close to the 5 with the exception of
resolution. The 4 is a 600x1200 dpi machine. But the dynamic range is in the 3.4
range and it also has the lower transparancy tray. It costs about $600 vs. the
5's $2200 price tag. Even though I;ve never used either machine, it would seem
like at 300 dpi they should produce close to identicle results.
> Stan
>
--
Come visit my site for information on digital photography and other
interesting topics. The site isn't finished yet, but I'm working on it.
http://www.greer.simplenet.com
Mike Greer
pbri...@wco.com
> First, we need to find a scanner that can handle 8 x 10 inch glass negatives.
> I'm concerned that the thickness of the glass may make it difficult to use a
> standard scanner. Any suggestions on a scanner that would be good for this
> purpose?
>
If you scan the plates emulsion side down, the thickness will not be an
issue. If this yeilds a back to front image it is very easy to batch
process the files to correct this.
> Second, we're puzzling over the resolution and compression we should use for
> scanning and storing these negatives. If I've done the math right, scanning 8
> x 10 inches at 300dpi and 8 bits of grayscale (these are black and white)
> results in a 72MB image. Times 100,000 images, that's a lot of CD-ROMs!!!
>
Maybe DVD would be a better medium. Even on CDROM the resulting archive is
a lot more compact than the originals. I think one of the two is the
natural choice, offering the greatest longevity now available.
****************************************************************
Patrick Brinton pbri...@wco.com
vox: (707) 874 1898
Consultant Generalist
****************************************************************
Richard Moore
>
> First, we need to find a scanner that can handle 8 x 10 inch glass negatives.
> I'm concerned that the thickness of the glass may make it difficult to use a
> standard scanner. Any suggestions on a scanner that would be good for this
> purpose?
You want a good flatbed scanner with a transparency back-light. The
thickness of the glass would probably make it necessary to adapt the
hinge of the top, but it shouldn't be that difficult. I use a Agfa Arcus
II to scan 4 x 5 and 8 x 10 negatives with very good results when
printed offset lithography. Agfa makes a more expensive model with a
drawer for scanning transparent media that would be even better for your
needs. I believe they sell for around $4,000 and the Agfa software, used
with Adobe Photoshop is excellent. Obviously a drum scanner woud not
work.
>
> Second, we're puzzling over the resolution and compression we should use for
> scanning and storing these negatives. If I've done the math right, scanning 8
> x 10 inches at 300dpi and 8 bits of grayscale (these are black and white)
> results in a 72MB image. Times 100,000 images, that's a lot of CD-ROMs!!!
>> Since preservation is our prime goal, we don't want to lose noticeable detail
> in the digitization process. Any suggestions on what scanning resolution, and
> compression format, we should consider to reduce storage size while not losing
> noticeable quality?
300 dpi seems to be the standard for photo-CD images intended for fine
offset printing reproduction at 150 lpi. Having the images at this
resolution gives you the option of selling some as stock photos to maybe
cover some of your costs. If the intent is to be able to produce new
film negs for continuous tone, the resolution is considerably higher.
For monitor display only the resolution for storage can be 72 dpi, I
believe.
Jeremy Rowe, head of Media Development with Information Technology at
Arizona State University, and an authority on 19th century Arizona
photography, was involved in a similar project and might be able to
share his methods with you.
Sounds like fun.
Richard Moore
Greg Finn
> I don't think this is not on topic. Although many people
> associate digital photography with unnatural alteration
> of images, it is becoming a great tool to simulate the
> dark-room process of enlarging and chemical printing, and
> it has significant application in large format photography.
> For example, EverColor prints, based on digital technology,
> are superb quality and long lasting. So the question is not
> "if" but "when" digital technology becomes a very popular
> way of printing?
A laser enlarger prints digital images on photographic paper more than
ten times as fast as your high-end liquid ink printers can image.
Outfits like Durst have already augmented their high-end enlarger
lines with these (400 ppi continuous tone, twenty 40x50 inch
prints/hour). The issue of whether or not the final image has
photographic quality and archival character or not is pretty much a
moot one here. With fewer moving parts, better quality and much, much
greater speed, as against the cost of wet processing, this thing is a
real horse race. Some labs that do murals are already getting rid of
their Iris printers and buying these because of the speed issue alone.
Will a digital enlarger make its way into home darkrooms? I don't see
why not. You could manufacture one as a color head add-on right now.
The market now would be small, but I wouldn't be surprised to see it
happen in a year or two. I'm not an emulsion guru, but I don't see
why you couldn't use these for B&W also.
Consider dodging and burning. Ansel Adams used an array of small
lights for his 8x10 enlarger head that he could individually turn
off/on for area dodge/burn. Same thing here, but a lot better. You
could do whatever you wanted. All you would have to do is describe
the regions and the amount of dodge/burn desired. Edge feathering
would be automatic. Adjustments to the color emulsion layers for
drift as exposure was varied could also be accounted for to a degree I
suspect ... but I'm out of my water here.
> If your goal is preservation, you need to produce a digital
> master copy of each image. For the purpose of preserving every
> possible detail from your physical medium, the resolution of
> the digital master copy file should be matched to the maximum
> resolution of your glass negatives. For example, assuming that
> your glass negatives have maximum resolution of 20 lpm
> (lines per millimeter) then you need to scan that image at
> (20 x 2 x 25.4 = 1016) dpi (dots per inch) to achieve
> one-to-one correspondence between digital and physical dots.
>
> A master copy file with 1000 dpi will give you ((8 x 1000) x
> (10 x 1000) = 80,000,000 bytes) or 76.29 MB file of gray scale
> image. Considering CD Rom capacity of about 650 MB, you could
> store only about 8 image files per CD (TIFF format). I realize,
> this would take a very large amount of resources, but I don't
> believe that each of the 200,000 negatives is a masterpiece
> worth such effort. I would first evaluate all images and
> select those with any artistic values, and scan them with the
> highest resolution. The rest that have some historical value
> could be scanned with lower resolution.
This is a large-scale archival problem. People costs will be a huge
factor. A man year is about 100,000 minutes. If a plate requires one
minute to scan and archive, that's two man years. Two minutes is four
man years. Triaging the images won't affect that cost much at all and
could conceivably raise it. The people budget here could be $100K and
could easily rise to be $300K.
You need to estimate the time per plate to get the people costs
ballparked. You might find that the people costs equal or surpass the
media costs.
net...@e-z.net
$4600, with memory for 12x18 page.
$8 per sheet (8x11) for paper and ribbon
--
Netcam Digital Photography, Less expensive than Prints/Scanning Method
and faster too! Web Page, E-mail, Property, Online Catalogue, Slide show
to Video.. Come see our Gallery at http://www.e-z.net/~netcam
Alexander Lee
folks
>are the most likely ones I can think of to have ideas on this topic.
>My company, NCR Corporation, has a huge historical archive. Among other
>things, we have what we believe is the world's largest collection of 8x10
>glass negatives (over 100,000 of them). We also have 100,000 or so
>hand-tinted glass magic-lantern slides.
>We've recently entered into an agreement with the local historical society
to
>make this archive accessible to the public. One of our first objectives is
to
>preserve these negatives and slides by converting them to digital format.
>That brings me to my questions.
negatives.
>I'm concerned that the thickness of the glass may make it difficult to use
a
>standard scanner. Any suggestions on a scanner that would be good for this
>purpose?
for
>scanning and storing these negatives. If I've done the math right,
scanning 8
>x 10 inches at 300dpi and 8 bits of grayscale (these are black and white)
>results in a 72MB image. Times 100,000 images, that's a lot of CD-ROMs!!!
>Since preservation is our prime goal, we don't want to lose noticeable
detail
>in the digitization process. Any suggestions on what scanning resolution,
and
>compression format, we should consider to reduce storage size while not
losing
>noticeable quality?
>area) magic-lantern slides, which have the added challenges of being in
color,
>and of having the emulsion sandwiched between two pieces of glass, so the
>image will be elevated a bit from the scanner bed.
>Thanks for any insight anyone can offer.
>John Ackermann
>john.ac...@daytonOH.ncr.com
Unfortunately, you we don't know what a ballpark figure for your project is.
I realize that it may be hard to come up with a figure at this time, and
even if you do have one, it would be hard to justify posting it on the net.
I would suggest that you contact Kodak's Eastman house or and see if they
have any advice or if they had previous projects to archive lantern slides
or glass negatives. Contact NASA for the image storage issues, as they have
had extensive experience in more than the quantity of image data you are
looking at. They would probably better be able to answer with direct
experience.
The scanning would also need to be done by someone with experience as an
archivist or graphic artist, as scans frequently need some tweaking from
image to image to get the best quality for differing contrast, density, etc.
Especially colour scans which need to have the colour fidelity checked.
As someone else mentioned, prioritizing what negatives and slides you need
to have maximum resolution of, and those that are acceptable with lower
resolutions is a priority, as scanning all of the images at a full 1000dpi
would need almost a terabyte of data storage or more. And then you have the
lantern slides. Not something to be put onto multiple CD's. As well, if
you are looking for archival storage, there are questions as to the long
term usability of CD Roms as well (not in the next couple of years, but 50,
100 years down the line if you are talking archival.)
If you are planning to make this readily accessible to the public, I would
suggest you consider using the flashpix format, as only a few people
accessing the any other format of the full resolution images would put a
strain on any network quickly. There will be extra storage space needed,
but the advantage in delivery should greatly compensate. As well, if you
were considering it, you would not need to store and track thumbnails for
browsing, medium resolution for quick display, and full resolution for
archival purposes.
For storage, I'd suggest for the moment, if you have the budget, some file
server with a large RAID drive. Pulling the image off of the hard drive is
much faster than from a cd, as well, you won't need a cd changer in order to
have them all online at one time. Perhaps you could get some large computer
company to donate some equipment? I work with Hewlett Packard's extensively
at my job, and using RAID5, you should easily be able to get a rackmount LX
server with half a terabyte (possibly 1 full terabyte, but I would be
hesitant about the performance) so getting one or two of them should fit
your bill. Other vendors have comparable products, I'm just using equipment
I am familiar with for an example. Later, you could burn them onto DVD's
when the technology comes down in price and they shake everything out for
distribution if needed. You should also consider getting some sort of tape
backup to insure that should some disaster happen, all the work does not go
down the tubes. As well, you would have to refresh the tapes every few
years to insure there is no data corruption from storage.
Lastly, you have not touched on image management. With about 100,000
images, it would be hell to find anything. Several organizations that deal
with large amount if imaged documents that need to be indexed and quickly
searchable and viewable, such as American Express, stock image libraries,
large magazines with many staff photographers such as National Geographic,
law firms, etc have customized applications for this. You may want to poll
them for recommendations for vendors, or pointers in that direction.
This is a project that can easily span several years and require the
attention of several full time staff. Asking people with previous
experience in some of the areas, as mentioned before like Kodak's Eastman
house, NASA, National Geographic, should be high on your list of priorities
to get a general idea of the scope and costs of such a project. Contacting
companies to donate equipment and expertise for a charitable cause for some
publicity would probably help the bottom line a bit as well.
Alex
zo...@panix.com
Vince
> Since preservation is our prime goal, we don't want to lose noticeable detail
> in the digitization process. Any suggestions on what scanning resolution, and
> compression format, we should consider to reduce storage size while not losing
> noticeable quality?
Buy at the very least a 10 bits/color scanner, 12bit preferred since the
negative can have dark areas or very subtil light ones that a 8bit/color
will miss.
In fat, if budget is not an issue by 2 scanners an 3 G3 Power Macintoshs
and ethernet 100baseT network:
Like that you could prepare the scanning of one plate on unit one (Mac 1
+ scanner 1) while the other unit (scanner2 + scanner 2 churns on the
other. You'll save a lot of time doing that, increasing your
productivity a lot.
Why a Third MAC (without scanner) ? It would be used as a server and
will AUTOMATICALLY compress the images, name them, and store them.
Divide the server HD in partitions, creating 3 of 650MB ones. And attach
a CD-R recorder to that mac.
Why G3 Macs ! Because they're the fastest for taht us, and they have the
wondegfull applescript that will automate the compression, the naming,
the storage of the files.
Of course if you can dedicate a forth mac to cd-burning it will be even
better.
Since your purpose is archiving, that means you don't want to ever
re-scan a plate, so I would definetely consider 300dpi (if you need a
modest enlargment capability).
(BTW your maths are off 8*300*10*300= 7.2 MB not 72 !)
But you have to test it, if Ed is right about the amount of film-grain
of the plate go for 150dpi rather than 100 because with scanners you
have to scan by multiple of two of their max optical resolution. If it's
a 600dpi scanner (wich is the least since I don't know a 300dpi scanner
with the required quality) you have to scan at 600, 300, 150, 75. If not
the scanner will interpolate wich is bad. If you eally want 100dpi, just
resize in phostoshop ! But be warned that you won't be able to print
that at a 8*10 size. For that you need 200dpi at least (300dpi
preferred).
Anyway, this is a DAUNTING task so you have to have the right tools,
fast g3 macs with at least 64Mb of memory are NEEDED !
John Ackermann
>I realize this question isn't quite on topic for these groups, but you folks
>are the most likely ones I can think of to have ideas on this topic.
>
>My company, NCR Corporation, has a huge historical archive. Among other
>things, we have what we believe is the world's largest collection of 8x10
>glass negatives (over 100,000 of them). We also have 100,000 or so
>hand-tinted glass magic-lantern slides.
etc.
I just wanted to thank everyone who's replied to my posting in the last 24
hours. I've been on the net for nearly ten years, and I've have more -- and
more useful -- responses to this question than anything else I've ever posted.
Thanks!!!
John Ackermann
john.ac...@daytonOH.ncr.com
Dave Martindale
>First, we need to find a scanner that can handle 8 x 10 inch glass negatives.
>I'm concerned that the thickness of the glass may make it difficult to use a
>standard scanner. Any suggestions on a scanner that would be good for this
>purpose?
One point that may be obvious, but nobody else seems to have mentioned:
you need a scanner with a transparency adapter. This is a device that
goes on top of the scanner, replacing the pressure plate. It has a
light source that moves along the scanner's glass plate in sync with
the CCD that is travelling under the plate. In this way, you get data
from the scanner that represents the amount of light transmitted from
the plate.
If you try to scan film or glass plates without a transparency adapter,
you will get some sort of image. But in this case, the light has made
*two* passes through the image, doubling the apparent density. Worse,
those two passes didn't go through exactly the same part of the image,
and one of those parts wasn't in focus, so you've blurred image detail.
This is even worse for glass plate than film, because the light has to
go well past the emulsion before being reflected by the white backing.
Depending on how tightly the illumination system focuses the light on
the area being scanned, and the angle of the illuminating beam, most
of the light may not even return to the area of the CCD.
So whatever you choose, it needs to be designed for scanning transparent
materials.
Dave
Otto Giesenfeld
>The only file format that you should store master copies are
>the ones that use lossless compression algorithm for example
>TIFF, PNG or BMP. Do not use JPEG format because it uses a
>lossy compression algorithm, and, depending on level of
>compression, it will corrupt some of the pixels permanently.
I don't agree with this statement. When scanning a glass plate,
you will necessarily lose a certain amount of information, as
the scanned pixels will be averages of several emulsion
grains. In comparison, the additional loss of information due
to a modest JPEG compression (say, factor one to five) would
not be terrifyingly large.
Quite on the contrary, with a given amount of storage space
at hand, JPEG compression will give you a _better_ image
quality than LZW-compressed TIFF or other lossless formats.
This is because you can afford to scan at twice the
resolution without needing more space.
Just be sure to select a moderate compression level. I suggest
it is set so that the average image gets approximately 1:5
compression for black and white, and 1:10-1:15 for color.
Applying artificial sharpening to the scanned image is not
necessarily a good idea, since it will make the JPEG
compression worse. And the sharpening could always be done
later on when retrieving the image.
Good luck, it sounds like an interesting project!
Otto Giesenfeld
Andrzej Irla
>
> In <35EB1F76...@tandem.com> Andrzej Irla <Andre...@tandem.com> writes:
>
> >The only file format that you should store master copies are
> >the ones that use lossless compression algorithm for example
> >TIFF, PNG or BMP. Do not use JPEG format because it uses a
> >lossy compression algorithm, and, depending on level of
> >compression, it will corrupt some of the pixels permanently.
>
> I don't agree with this statement. When scanning a glass plate,
> you will necessarily lose a certain amount of information, as
> the scanned pixels will be averages of several emulsion
> grains. In comparison, the additional loss of information due
> to a modest JPEG compression (say, factor one to five) would
> not be terrifyingly large.
>
I think, you didn't read my message very carefully. In paragraph
two, I clearly stated, "the resolution of the digital master copy
file should be matched to the maximum resolution of your glass
negatives... to achieve one-to-one correspondence between digital
and physical dots." So, if you follow that directions, your
scanner's color sensor will sample an area smaller or equal to the
smallest distinguishable detail on the physical medium and won't
have to average them.
> Quite on the contrary, with a given amount of storage space
> at hand, JPEG compression will give you a _better_ image
> quality than LZW-compressed TIFF or other lossless formats.
> This is because you can afford to scan at twice the
> resolution without needing more space.
>
No, your assumption is wrong. I hope, you don't think that
increasing scanning resolution always gives better image
quality. It is only true up to the point when the scanning
resolution matches the negative resolution. Further increase
of density would not bring a bit of improvement to the digital
image. It is not hard to notice that there is a limitation in
the physical medium, and even the best scanner will not do
any magic. Since John Ackermann's goal is preservation without
losing noticeable detail, I think that JPEG format would be a
very poor choice for preserving images possessing some artistic
values. You never know what you are going to do with these
files in the future. Existence of defective pixels may limit
their use for making larger prints. On the other hand, images
with only historical values could be stored in JPEG only for
the sake of saving disk space.
Helge Nareid
wrote:
>John Ackermann <john.ac...@daytonOH.ncr.com> wrote:
>
>> Since preservation is our prime goal, we don't want to lose noticeable detail
>> in the digitization process. Any suggestions on what scanning resolution, and
>> compression format, we should consider to reduce storage size while not losing
>> noticeable quality?
>
>Buy at the very least a 10 bits/color scanner, 12bit preferred since the
>negative can have dark areas or very subtil light ones that a 8bit/color
>will miss.
This is a very good point. The dynamic range of these plates is likely
to be much wider than modern negatives, for two reasons. Firstly, they
were most likely exposed without an exposure meter, and will vary
considerably in density for that reason. Secondly, they are probably
developed for contact-printing on printing-out paper, and will
therefore be more dense and more contrasty than the modern norm. A
fairly large dynamic range is therefore needed to get the optimum
tonality from these plates.
--
- Helge Nareid
Nordmann i utlendighet, Aberdeen, Scotland
Jim Bancroft
John Ackermann wrote in message ...
>My company, NCR Corporation, has a huge historical archive. Among other
>things, we have what we believe is the world's largest collection of 8x10
>glass negatives (over 100,000 of them). We also have 100,000 or so
>hand-tinted glass magic-lantern slides.
John -
Another solution might be to use a copy stand setup with a light box and a
35 mm camera with a bulk film back to transcribe the originals. A single
operator might be able to transcribe about 1-2000 per day (very boring
work). The 35mm film can be scanned using a commercial 35mm scanner with a
roll film feeder. You may be able to farm this part to a service bureau.
You might want to consider tape as a storage medium. Access is not super
fast but capacities upto terabytes make it possible to have large numbers of
images online at a moderate cost.
jim
Otto Giesenfeld
>I think, you didn't read my message very carefully. In paragraph
>two, I clearly stated, "the resolution of the digital master copy
>file should be matched to the maximum resolution of your glass
>negatives... to achieve one-to-one correspondence between digital
>and physical dots." So, if you follow that directions, your
>scanner's color sensor will sample an area smaller or equal to the
>smallest distinguishable detail on the physical medium and won't
>have to average them.
Perhaps I should have expressed myself more clearly. My main point
was that when converting an image from a grain-based medium like
glass plates to a pixel-based one like digital files, there is no
well-definied optimum or "matching" resolution.
Until the point where you are able to scan at a resolution (10 000
dpi or more?) that will capture the shape of each individual emulsion
grain, you will always be able to get _some_ more image detail by
increasing resolution.
Therefore, I still think that scanning at a high resolution and
using some JPEG compression is a good idea. Some people seem to
be scared by the fact that JPEG is a "lossy" compression, but
the loss is actually very limited if the compression rate is
modest. Compare it to contact printing the plates on modern 8x10
sheet film. Obviously, the detail captured would satisfy the
most demanding archivist. But the fact is that 8x10 sheet film
is a lossy storage medium, just like JPEG-compressed computer
files. It is just that the loss with sheet film storage is so
marginal compared to the storage capacity that we can neglect it.
Otto Giesenfeld
Dave Martindale
>Secondly, they are probably
>developed for contact-printing on printing-out paper, and will
>therefore be more dense and more contrasty than the modern norm.
A very interesting point. Do you have any idea what sort of gamma
such negatives might have been developed to? (Current practice is
about 0.6 for colour negatives, and up to about 0.65 for B&W).
Dave
Dave Martindale
>Another solution might be to use a copy stand setup with a light box and a
>35 mm camera with a bulk film back to transcribe the originals. A single
>operator might be able to transcribe about 1-2000 per day (very boring
>work). The 35mm film can be scanned using a commercial 35mm scanner with a
>roll film feeder. You may be able to farm this part to a service bureau.
This would work if you're happy with the resolution that can be achieved
by modern-day B&W film in 35mm. But it's fairly safe to assume that even
though the 8x10 negatives use older, much coarser-grained emulsions, they
still have more detail than a 35mm negative can resolve.
Dave
Jim Bancroft
I don't believe that they will be able to achieve a full record of the 8X10
glass negatives with any current technology. Even if it was possible the
storage requirements would be prohibitive.
My suggestion was aimed at a practical solution to provide access to their
archives using simple conventional methods.
jim
Dave Martindale wrote in message <6sk0v1$543$1...@columbia.cs.ubc.ca>...
Mike McDonald
"Jim Bancroft" <banc...@home.com> writes:
> Dave-
>
> I don't believe that they will be able to achieve a full record of the 8X10
> glass negatives with any current technology. Even if it was possible the
> storage requirements would be prohibitive.
Prohibitive? How much space do you think the 100,000 glass negatives takes
up now? Even if it took 3 CDs for each image, that'd still take up less space
than the glass negatives. (8x10 at 2Kdpi -> 320M pixels * 4bytes/pixel -> ~3
CDs)
Mike McDonald
mik...@mikemac.com
Jim Bancroft
Yes. But now you have both.
jim
Mike McDonald wrote in message ...
...
Alexander Lee
Dave Martindale wrote in message <6sk0v1$543$1...@columbia.cs.ubc.ca>...
>"Jim Bancroft" <banc...@home.com> writes:
>>Another solution might be to use a copy stand setup with a light box and a
>>35 mm camera with a bulk film back to transcribe the originals. A single
>>operator might be able to transcribe about 1-2000 per day (very boring
>>work). The 35mm film can be scanned using a commercial 35mm scanner with
a
>>roll film feeder. You may be able to farm this part to a service bureau.
>This would work if you're happy with the resolution that can be achieved
>by modern-day B&W film in 35mm. But it's fairly safe to assume that even
>though the 8x10 negatives use older, much coarser-grained emulsions, they
>still have more detail than a 35mm negative can resolve.
> Dave
As these are glass plate negatives, they were made from a wet plate process.
The wet plate process makes a negative that is virtually grainless, so the
limiting factor for resolution will be the darkroom skill, lens and
photographic ability of the photographer. View Camera magazine had a decent
article that briefly described the wet plate process in its July/August
issue, and information should be available on the web. Just search for wet
plate photography or collodion.
Alex
zo...@panix.com
Max Buten
>...>
>>My company, NCR Corporation, has a huge historical archive. Among other
>>things, we have what we believe is the world's largest collection of 8x10
>>glass negatives (over 100,000 of them). We also have 100,000 or so
>>hand-tinted glass magic-lantern slides.
Not that it's any of my business, but I'm curious. What's on these
big glass negatives? And why does NCR have so many?
Max
pbri...@wco.com
>As these are glass plate negatives, they were made from a wet plate process.
>The wet plate process makes a negative that is virtually grainless, so the
>limiting factor for resolution will be the darkroom skill, lens and
>photographic ability of the photographer. View Camera magazine had a decent
>article that briefly described the wet plate process in its July/August
>issue, and information should be available on the web. Just search for wet
>plate photography or collodion.
>
Are you sure about this? If my memory serves me, the wet plate process
involved exposing onto a plate that was wet. I used a 4 x 5 view camera in
England back in 1962 or so that had holders for glass plates. I mostly
used cut sheet film in it, which meant using different holders, but I was
able to purchase and use glass plates with dry emulsion, which I did on
several occasions.
This is not to say that these particular plates were not made with the wet
plate process, but it is my impression that this was in fact a fairly
early process supplanted by dry emulsion glass plates some time before
WW2. However I do not claim to be an expert in the field, and am relying
on memory from some time back.
Jean-David Beyer
> Dave Martindale wrote in message <6sk0v1$543$1...@columbia.cs.ubc.ca>...
> >"Jim Bancroft" <banc...@home.com> writes:
> >>Another solution might be to use a copy stand setup with a light box and a
> >>35 mm camera with a bulk film back to transcribe the originals. A single
> >>operator might be able to transcribe about 1-2000 per day (very boring
> >>work). The 35mm film can be scanned using a commercial 35mm scanner with
> a
> >>roll film feeder. You may be able to farm this part to a service bureau.
> >This would work if you're happy with the resolution that can be achieved
> >by modern-day B&W film in 35mm. But it's fairly safe to assume that even
> >though the 8x10 negatives use older, much coarser-grained emulsions, they
> >still have more detail than a 35mm negative can resolve.
> > Dave
>
> As these are glass plate negatives, they were made from a wet plate process.
Jumping to conclusions there. George Eastman made his fortune from making glass
dry-plates. While I have no statistics to back this up, I would imagine that,
except for the graphics arts industry perhaps, more dry plates were made than
wet plates from the beginning of the world (about 1839) until today.
>
> The wet plate process makes a negative that is virtually grainless, so the
> limiting factor for resolution will be the darkroom skill, lens and
> photographic ability of the photographer. View Camera magazine had a decent
> article that briefly described the wet plate process in its July/August
> issue, and information should be available on the web. Just search for wet
> plate photography or collodion.
>
> Alex
> zo...@panix.com
--
Windows is not the answer. Windows is the question. The answer is No.
Jean-David Beyer
Shrewsbury, New Jersey
Helge Nareid
wrote:
>As these are glass plate negatives, they were made from a wet plate process.
Not necessarily. Dry plates were introduced around 1880, and rapidly
superseded the wet-plate process (except for specialized applications -
I've been told by a reliable source that wet plates were used as late as
the 1960s for some applications).
Dry plates were quite commonly used for professional work until fairly
recently - the various film bases used in the first half of the century
were unsatisfactory for demanding applications, particularly regarding
dimensional stability. It was only when the polyester (Estar) base was
introduced that sheet film took over (nearly) completely. I believe the
Estar base was introduced sometime around 1950.
As a matter of fact you can still buy dry plates today - apart from a
number of highly specialized emulsions for applications such as
holography and high-resolution photolitography, I believe that T-Max 100
can still be supplied on glass plates.
--
Helge Nareid - Nordmann i utlendighet
University of Aberdeen, Department of Engineering
Laser and Optical Engineering Group
http://wwwcad.eng.abdn.ac.uk/~eng529/
Helge Nareid
Sorry - I don't have any quantitative data. I've printed a few old glass
plates myself, so I've seen it, but I don't have extensive first-hand
experience. A couple of friends of mine have worked in the photographics
archives of a museum printing old plates, and they have told me
something about the difficulties in getting a satisfactory print from
them.
John Ackermann
#def Corporate_Mouthpiece_Mode ON
NCR's founder, John H. Patterson, was arguably the creator of modern sales and
marketing techniques. He wanted to have photographs his product (cash
registers) in use and so he hired legions of photographers to take pictures of
stores, etc. where NCR products were installed. And, he told the
photographers that he wanted people in the pictures, so these aren't just
still life shots -- they're real slices of life, with customers, employees,
and passers-by.
Since NCR rapidly became an international company (our first non-US locations
opened before the turn of the century), the pictures come from all over the
world. And, along the way, lots of other interesting images were created. We
have pictures of the San Fransisco earthquake, for example. And celebrities
visiting NCR facilities. And lots and lots of pictures of NCR people at work
and play, both in Dayton and around the world. The picture-taking began in
the 1890s, and continued at full force through the 1920s, and at a
tapering-off pace through the 70s (since then, we have a more normal
corporate archive).
On top of it all, Patterson was something of a pack rat, and that rubbed off
on the company, so there are lots of other fascinating items in the collection
-- over 3,000,000 items in all. We have about 3,000 reels of motion picture
film (which we're also rushing to preserve via transfer to video). We have a
few tiny bits of the original Wright Brothers Flyer of 1903 (Patterson and
Orville and Wilbur Wright were buds). We have hundreds and hundreds of early
cash registers and accounting machines. We have every engineer's notebook
from the early days. We have the world's first fully solid-state computer,
manufactured by NCR in the '50s (don't have my reference material handy to
provide the actual date). We have... well, you get the picture.
For years, we've been trying to figure out how to both preserve this
collection and make it accessible to the public, given that the company had
lots of other problems to focus on (acquisition by AT&T, spin-off by AT&T,
working to make it on our own, etc.). This summer we came up with an
arrangement with the Montgomery County Historical Society that we believe will
accomplish this purpose, and the digitizing project is part of that.
#def Corporate_Mouthpiece_Mode OFF
John Ackermann
john.ac...@daytonOH.ncr.com
Mike McDonald
"Jim Bancroft" <banc...@home.com> writes:
> Mike McDonald wrote in message ...
> ...
>> Prohibitive? How much space do you think the 100,000 glass negatives
> takes
>>up now? Even if it took 3 CDs for each image, that'd still take up less
> space
>>than the glass negatives. (8x10 at 2Kdpi -> 320M pixels * 4bytes/pixel ->
> ~3
>>CDs)
>>
>
> Mike-
>
> Yes. But now you have both.
>
> jim
Actually, it's worse than that. They had better make multiple copies of the
digital copies and store them in separate locations. Having a fire wipe out
both the original glass negs and the digital ones would be a crime.
Mike McDonald
mik...@mikemac.com
Roderick Stewart
> >Not that it's any of my business, but I'm curious. What's on these
> >big glass negatives? And why does NCR have so many?
> >Max
>
> #def Corporate_Mouthpiece_Mode ON
>
> NCR's founder, John H. Patterson, was arguably the creator of modern sales and
> marketing techniques. He wanted to have photographs his product (cash
> registers) in use and so he hired legions of photographers to take pictures of
> stores, etc. where NCR products were installed. And, he told the
> photographers that he wanted people in the pictures, so these aren't just
> still life shots -- they're real slices of life, with customers, employees,
> and passers-by.
>
Sounds like you have a historical goldmine there. You obviously have the right
idea in wanting to make the best possible scans for posterity, but whatever you
do, don't destroy the originals. (See "The size of Thoughts" by Nicholson Baker).
Rod.
Chris Buechner
>
> As a matter of fact you can still buy dry plates today - apart from a
> number of highly specialized emulsions for applications such as
> holography and high-resolution photolitography, I believe that T-Max 100
> can still be supplied on glass plates.
>
I believe glass plates are still used for scientific purposes, where
distance measurements have to be taken of the neg. This can't be done
100% accurately with estar based negs, as they have some degree of
distortion, even heat and cold will affect them. Glass on the other hand
will not distort under normal conditions.
My 2c worth.
See ya
--
Chris Buechner
WILDfire [ photography ]
Sydney, Australia
wild...@tpgi.com.au
"You don't take a photograph, you make it" Ansel Adams
Alexander Lee
I guess everyone knows what happens when you assume....
Alex
zo...@panix.com
Helge Nareid wrote in message <35ee5fcb...@newshost.abdn.ac.uk>...
>On Wed, 2 Sep 1998 18:17:22 -0400, "Alexander Lee" <zo...@panix.com>
>wrote:
>>As these are glass plate negatives, they were made from a wet plate
process.
>Not necessarily. Dry plates were introduced around 1880, and rapidly
>superseded the wet-plate process (except for specialized applications -
>I've been told by a reliable source that wet plates were used as late as
>the 1960s for some applications).
>Dry plates were quite commonly used for professional work until fairly
>recently - the various film bases used in the first half of the century
>were unsatisfactory for demanding applications, particularly regarding
>dimensional stability. It was only when the polyester (Estar) base was
>introduced that sheet film took over (nearly) completely. I believe the
>Estar base was introduced sometime around 1950.
>As a matter of fact you can still buy dry plates today - apart from a
>number of highly specialized emulsions for applications such as
>holography and high-resolution photolitography, I believe that T-Max 100
>can still be supplied on glass plates.
Sandor Mathe
: Helge Nareid wrote:
:
: >
: > As a matter of fact you can still buy dry plates today - apart from a
: > number of highly specialized emulsions for applications such as
: > holography and high-resolution photolitography, I believe that T-Max 100
: > can still be supplied on glass plates.
:
: I believe glass plates are still used for scientific purposes, where
: distance measurements have to be taken of the neg. This can't be done
: 100% accurately with estar based negs, as they have some degree of
: distortion, even heat and cold will affect them. Glass on the other hand
: will not distort under normal conditions.
What kind of glass is used?
If its normal window pane (soda) glass then it does have a thermal expansion
that may be important in a science experiment, not to mention the green cast.
If its boro-silicate ("Corning wear") as is used in telescope mirrors the
thermal expansion is much less.
I doubt they use pure quartz glass.
--
Sandor Mathe
san...@ca.ibm.com
Peter Jones
Interesting project. I suggest that you visit Harvard Library's
Preservation site. While it is a bit old ('95/'96) the task you are
describing is closer to what Harvard is doing than your typical
newgroup contributor.
http://preserve.harvard.edu/resources/digitization/scanning.html
It might be helpful to tackle the problem from the other end.
Consider what you are going to do with the scans once you have them,
and use this information to rule out various digitizing options. For
example, if you want to provide this to the world on web pages, you
don't need very high resolution. On the other hand, if you want to
preserve every last bit of information, then at the very least this
rules out options like digital cameras/photoCD/film.
In addition, a project of this considerable size may require the
creation of a group to tackle the problem. Once they have finished,
what should they do next? Perhaps NCR could allow other companies to
hire their expertise.
Peter
john.ac...@daytonOH.ncr.com (John Ackermann) wrote:
>In article <35edc6e2.517852298@news>, maxb...@home.com (Max Buten) wrote:
>>>John Ackermann wrote in message ...
>>>...>
>>>>My company, NCR Corporation, has a huge historical archive. Among other
>>>>things, we have what we believe is the world's largest collection of 8x10
>>>>glass negatives (over 100,000 of them). We also have 100,000 or so
>>>>hand-tinted glass magic-lantern slides.
>>
>>Not that it's any of my business, but I'm curious. What's on these
>>big glass negatives? And why does NCR have so many?
>>Max
>
>#def Corporate_Mouthpiece_Mode ON
>
>NCR's founder, John H. Patterson, was arguably the creator of modern sales and
>marketing techniques. He wanted to have photographs his product (cash
>registers) in use and so he hired legions of photographers to take pictures of
>stores, etc. where NCR products were installed. And, he told the
>photographers that he wanted people in the pictures, so these aren't just
>still life shots -- they're real slices of life, with customers, employees,
>and passers-by.
>
Gregory Benjamin
superior to polyester because it is affected only by temperature
and not by temperature and humidity as polyester is. The dimensional
response of polyester-based film to humidity is non-linear and
exhibits hysteresis. That is, if you elevate the humidity, the
size increases, but when you reduce the humidity back to the original
value the size stays large. To get the film back to the original size
you have to drive the moisture out of it by using elevated temperatures
at lower than normal humidity.
Glass plates don't have this problem. The size changes linearly with
temperature, and there is no hysteresis. To establish a specific
dimension, the plate is exposed at a particular temperature, and brought
to the same temperature when used. While specialized glasses like
Zerodur offer much lower (better) thermal expansion coefficents than
soda lime float glass, these materials are rarely required, and are
not offered as standard items by either Agfa or Kodak, the two
predominant photoplate suppliers.
Sandor Mathe wrote in message <6soqfv$16do$1...@tornews.torolab.ibm.com>...
Valburg
> With few exceptions, in many disciplines of photography like
> landscape or portrait, grain is an undesired feature. It
> is also one of the reasons, people use large format instead of
> 35mm. So, there is no reason to preserve the grain in a
> digital file as well.
What you say is true, in the context of photographic aesthetics, but I
can't help but wonder if the context of historical preservation mightn't
suggest otherwise. Is the subject of the archive the scene photographed
or the glass plate as an artifact? (rhetorical question)
Best,
Mitch Valburg
Andrzej Irla
> Perhaps I should have expressed myself more clearly. My main point
> was that when converting an image from a grain-based medium like
> glass plates to a pixel-based one like digital files, there is no
> well-definied optimum or "matching" resolution.
>
> Until the point where you are able to scan at a resolution (10 000
> dpi or more?) that will capture the shape of each individual emulsion
> grain, you will always be able to get _some_ more image detail by
> increasing resolution.
>
> Therefore, I still think that scanning at a high resolution and
> using some JPEG compression is a good idea. Some people seem to
> be scared by the fact that JPEG is a "lossy" compression, but
> the loss is actually very limited if the compression rate is
> modest. Compare it to contact printing the plates on modern 8x10
> sheet film. Obviously, the detail captured would satisfy the
> most demanding archivist. But the fact is that 8x10 sheet film
> is a lossy storage medium, just like JPEG-compressed computer
> files. It is just that the loss with sheet film storage is so
> marginal compared to the storage capacity that we can neglect it.
>
> Otto Giesenfeld
I think, this argument diverges from any realistic application.
I the digital transformation, the goal would be to preserve
only the effective image, not the grain structure of the physical
medium. That makes sense because digital pixels store only color,
or gray level information. Besides, the grain has also its minimum
diameter and pushing the scanning density further will only reveal
the crystal structure of silver grain. Would that improve image
quality or rather obscure it?
The effective image recorded on light sensitive film or glass
consist of collection of dots that strictly correspond to the
real image, as seen through the lens. The grain present on film
do not correspond to anything in the real image, and, therefore
it is not part of the effective image. It is only an unavoidable
characteristic feature of photographic medium.
With few exceptions, in many disciplines of photography like
landscape or portrait, grain is an undesired feature. It
is also one of the reasons, people use large format instead of
35mm. So, there is no reason to preserve the grain in a
digital file as well.
Since the goal would be to reproduce the effective image in a
digital format, and the resolution of an effective image
is limited by the lens/film combination, therefore, there is a
matching, scanning resolution that produces the optimum quality,
digital image. This optimum scanning resolution is simply equal
to the lens/film resolution for given f-stop. It is very easy
to find out what is that resolution without even going out of
the domain of this newsgroup. Simply, refer to Ch. Perez and
K. Thalmann's web site and calculate the average resolution
for a given f-stop. Their tests not only show the performance
of a particular lens, but also, it gives an idea of how much
quality can be expected from lens/film combination in real,
field situation.
The loss of data in JPEG format may be negligible in case of
contact print sizes, but it will show up when it comes to
print bigger formats. One thing we have to remember, like with
other things in nature, there is always a trade-off between
storage capacity and data accuracy. It is impossible to gain
one without loosing the other.
Regards
Andre Irla