Re: Characterize Digital Camera Color

[Ruey <tmi@xxxxxxxxxxxxx>]

James B. Davis wrote:

On Fri, 18 Dec 2009 12:17:29 -0700, Ruey <tmi@xxxxxxxxxxxxx> wrote/replied to:

  

Trevor, I'd bet that you would be able to tell the difference between a 
good dye transfer and a gum bichromate if you tried, and even tell that 
one more closely resembled what you see with your eyes. Just as it is 
possible for the human vision system to detect the difference between an 
out of focus image and one that is sharp, it is generally possible for 
most people to tell the difference between images whose color accurately 
reproduces a subject and one that abstracts color (excepting perhaps 
people whose red sensor overlaps their green one or the less frequent 
occurrence of blue overlaps green, or where reality just becomes 
incomprehensible). Rather than abuse my interest in obtaining accurate 
reproduction of color why not just keep silent if the posting does not 
interest you?
    

Ed, it is my understanding that all humans see colours very differently. I'm not
sure then how you intend to 'accurately reproduce' what the human eye sees?
Whose eyes? Yours? A panel of judges?

Is it then real art?

  

I haven't time to argue the case but instead would refer you to the classic text on color, "The Reproduction of Colour" by R.W.G. Hunt of Kodak Labs in Britain. The primary anomolies in what we perceive as color are a merging of data from red and green sensors in our eyes which we call red-green color blindness, and the much rarer merging of "blue-green" sensor data or blue-green color blindness. What is amazing is how much agreement does exist about what correct color is and if that were not true one imagines that Kodak, Land, Agfa, Ilford and a few other companies would not have made such extensive efforts at perfecting color photography.

As an R&D engineer I had the privilege of working with the first 2048x2048 color Trinitron. I designed the bitmap memory and went on to write all of the Phigs microcode for its graphics controller. We were teamed one-to-one with Sony engineers and had 6 months to design, build hardware, write microcode, UNIX software and test it out - an incredible project for its time. It was for an FAA console. Hughes, who I worked for, foolishly went to the FAA and asked if it would be OK if they bid a 2048x2048 display instead of a 1280x1024 one (the state of the art at the time). The FAA rewrote the RFP to require a 2048x2048 display and we lost our advantage. The FAA awarded the contract to IBM Government Systems who under bid, went on to make a complete mess of the FAA redesign and IBM eventually sold off Government Systems because of lawsuit for failure to perform. The FAA did say we had produced a technically superior system just a little more money than they wanted to spend. I know a fair amount about color reproduction and this is why I find the Russian Prokudin-Gorsky images so remarkable. I wish I could follow him around for a week or more and pick his brain. What an incredible person he must have been to have mastered the technical process and then travel around Russia to take so many fascinating images - the quintessential artist/scientist.

Ed Scott