Digital image to archival print

[Robert A. (Bob) Morris]

I'm not conversant with the requirements for such archival images, but
as someone who teaches computer science students about color I can
mention a few things about which I find few biologists are only
dimly aware: it is not at all trivial to assure that a digital image
will have a similar appearance when rendered on different media and
viewed under differing illuminations conditions. On the other hand,
for dried herbarium specimens, maybe nobody cares about color accuracy
anyway, in which case you can skip all my opinions except 1 and 2.

1. Use the highest imaging resolution you can afford, and use
lossless compression---or no compression. The highest quality jpeg
compression commonly implemented not lossless, except to human vision.
In the future, when you find you want to do digital image processing
by software instead of by eye, you will bless this advice.
Storage requirements are a red-herring. Disk space is approaching
$1/gigabyte.

2. Kodak has good tutorials and tips at
http://www.kodak.com/US/en/digital/dlc/

Among the important color issues are:

3. The range of colors reproducible by most printing
technologies (the "color gamut") is generally quite a bit less than
that of digital cameras or CRTs. Consequently, by definition, printing
software "distorts" the colors by mapping out-of-gamut image colors to
colors that can be produced by the printer. High-end image processing
software like Photoshop does this in standard ways, BUT those ways
depend on accurate color characterization (a technical term that may be
taken in its conversational sense) of each piece of the imaging chain:
the illuminant of the image acquisition system, the image acquisition
system itself, any compression used (e.g. jpeg) in storing the image,
the paper, inks, and printer used to print the image, and the
illuminant used to view the print. As an alternative to the
time consuming and technically demanding characterization processes
you should do one or preferably all of these things:
    a. Understand and rigidily follow the instructions about lighting
conditions recommended by the camera manufacturer.
    b. Photograph on a standard neutral gray card, such as is
available from Kodak.
    c. Photograph a standard color card with the specimen. This will
often make it possible for an expert to "invert" the color
distortions introduced and produce, under various conditions, a
satisfactory image. A reasonably satisfactory alternative is to
photograph the color card just against the background, and with the
same lighting conditions as the specimen(s) are imaged, and archive
that image. Take special care about the next recommendation:
     d. Carefully follow the manufacturer's recommendation about the
longevity of the lights you use and the color card. Care for both as
though they were as precious as your specimens. Lights change their
spectra as they age.

2. If you put your images on the web or distribute them to others,
learn to live with the fact that you have no control over the
reproduction conditions. Expect that scientific deductions based on
subtle color may have to take into account everything said above to be
meaningful. About half the people who look at your images will
complain they are washed out. About half will complain they are
unrealistically colorful. You will agree with each of them whose
facility you visit.

--Bob Morris


Michael Chamberland writes:
 > Date:         Thu, 10 May 2001 11:57:33 -0700
 > From: Michael Chamberland <chamb at u.arizona.edu>
 > To: TAXACOM at usobi.org
 > Subject:      Digital image to archival print
 >
 > As digital cameras have become more popular, we now have collectors who
 > would like to attach a photo to their herbarium specimen, but have only a
 > digital image for this purpose.  Any recommendations on how to transfer
 > from digital format to an archival image to add to a herbarium sheet?
 >
 > Thanks,
 >
 > Michael Chamberland
 > Assistant Curator
 > University of Arizona Herbarium (ARIZ)
 >