Re: Digital camera profiling
Re: Digital camera profiling
- Subject: Re: Digital camera profiling
- From: MARK MAIO <email@hidden>
- Date: Fri, 18 Apr 2003 14:41:27 -0400
First let me apologize for the length of this. I have been following the thread regarding
color profiling of digital cameras and thought I would add my personal experience. Let me
qualify what I am about to say by stating I am not, nor claim to be, an expert in the
digital field. Having been a member of this list for the past year, there are obviously
many contributors who are extremely knowledgeable and information I have gotten from them
has been very helpful. So what I am about to say might be way off base or already covered
by someone.
My 30 year background in photography is equally divided between medical/scientific
imaging in ophthalmology and my personal fine art work. In ophthalmic photography, we use
35 mm film (color and B&W), to do diagnostic imaging of the retina of the eye. These
images are used not only in research and diagnosis, but high magnification areas of each
frame are used to located almost microscopic sections of the retina while treatment with
laser is being done. Digital imaging in ophthalmology became commercially viable almost
20 years ago as the 35 mm film cameras we were using on specialized instrumentation to
image the retina, started being replaced by early B&W low resolution (512 X 512), CCD's.
Obviously detail and resolution did not compare to 35 mm film, but early adopters
impressed their patients with instant images and test results instead of the normal one
day turn around of film. Digital technology steadily improved over the years and
currently, about 20% of all imaging of the eye (both color and B&W), is being done
digitally.
I ran ophthalmic photography departments at three major medical centers during those
years, and was totally against digital because it did not (at that time), have the same
quality, consistency and detail as film. My belief was that patients deserved the highest
quality image for diagnosis and treatment and until digital could compare with film, I
was not going to adopt it. I provide this background to explain how I became involved
with the digital process and this issue of color calibration.
Seven years ago I left my university position in ophthalmology to devote more time to my
fine art. I did not quit my involvement in ophthalmic photography, but began teaching and
consulting in the field to pay the bills. At the same time, I became aware of the full
frame 35 mm sized Phillips CCD being used on digital backs for 2 1/4 cameras. I saw the
images these backs produced and was convinced the technology could be adapted to
ophthalmology. To make a long story somewhat shorter, I approached all the leading
manufactures of these backs with my idea and only one company, MegaVision, understood the
market potential and agreed to work with me. We developed the technology, partnered with
an ophthalmic company to start a new company and began selling the product.
In the process of developing the product, I learned a great deal about color calibration
of the CCD and how it related to what we did in color correcting our film. We knew that
in ophthalmology, photographing the same eye on one retinal camera with different brands
and types of color slide film ,will produce images of the retina with different color
casts. Commercial photographers have the same problem with film, whether in their studios
or outside. They solved it by calibrating their "system". Conducting tests using the same
camera/lens under controlled studio lighting systems or outside, they settled on
brands/types of film and had it processed at a quality controlled lab for consistency. By
doing this, they could come close to predicting that the color of the product they were
photographing would be the same on the final transparency they delivered to the client.
They also knew that using different lenses, camera systems, light sources or different
times of the day would affect the color balance of their film, so they tested each of the
components and used different levels of filtration to keep their system consistent to
what they wanted. Because most current CCD's start as monochrome and are made "color" by
applying a dye layer to each pixel (and these dyes used in the CCD production process
vary in color as much as different types and emulsion batches of film do), this same
problem of reproducing color correctly follows us into the digital realm.
At our company, we solved this problem in our ophthalmic digital imaging systems by
evaluating each full frame 35 mm size CCD we use and computing the variance in color
introduced in the production process of each of the six million pixels to a standardized
color target. We then create a "profile" CD which contains an algorithm that compansates
for the color shift and stays with that particular CCD/digital back. During installation
of our systems, this profile is loaded into our software and a color balance routine is
done on the entire system by photographing the same target the CCD was profiled for
through a software function in our capture program. The key here is color balancing the
"system", meaning not only the digital back and target, but just as important in
obtaining a color corrected system, the optics of the retinal camera and the color cast
produced by the flash tube. The initial reaction to these images was interesting. Since
most customers were switching from color film to digital, they were accustomed to the
exaggerated colors of whichever slide film they were using. It took a while to convince
them that our digital image was more "real" in appearance than the film they were used to
looking at. With a calibrated workflow and viewing stations, we were able to have
ophthalmologists and photographers view an image of the retina on their monitor that
matched the patient sitting in front of them. When you are using the color images to
monitor change over a period of time (and make diagnosis/treatment decisions based on
these photos), you want the most "real" representation possible. As in commercial
photography, should our customers still want the "appearance" of a certain film, we also
allow them to create custom developers to achieve this.
Obviously this profiling takes time and costs money. Our systems are hand built and
retail for $45,000.00. For that price you get a three day install by one of our
technicians, a calibrated workflow on all the hardware purchased from us and one year of
on site and phone technical support. Now don't get me wrong. I am not writing this to
suggest we need to spend 45 grand to get good digital images. We just have to understand
that when we, as commercial photographers, purchase one of the professional digital SLR's
in the $2,000 - $7,000 range, we get an unprofiled CCD and generic "developers" in the
software the company provides. Much like film, where we did exposure and color correction
tests on each type and emulsion batch we used, we need to do similar tests with each
digital back/camera we use, as there will be obvious differences in the color response
from the CCD used in your D100 bought a year ago and the one you bought last week. To
work around this, we buy additional software to help us find our "personal" developer and
run other tests to give us images that look the way we want them to. This takes a lot of
time, costs each of us money in the time spent doing it and in my opinion, is much more
complicated than what I did with film. To provide us with custom profiles of each CCD we
get in our digital SLR's, would currently add too much to the production process. My hope
is that in the near future, this type of information will be economically feasible to
include with each professional digital camera, making the process of getting the color
each of us wants, much easier.
Mark Maio
_______________________________________________
colorsync-users mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/colorsync-users
Do not post admin requests to the list. They will be ignored.