Digital BW, The Print

Alan Ansell wrote:
>  Hi Bruce,
>
> As I understand it, the basis for colour management - the CIE color space - was produced after extensive studies of the human visual system. Without going into a depth (there is a very full definition available on Wikipedia - search both "Color vision" and "CIE 1931 Color Space") and with respect to your experience in this field, I feel that is putting the cart a little before the horse.
>   
We've learned a bit since 1931. We've developed better instruments and 
can take better measurements. And we've learned more about what we 
should be measuring. We've found people with a fourth visual cone 
(orange). We've found that the cones we do have vary from person to 
person. One person has a "red" cone while the person sitting next to 
them has a "red-orange" cone. We've learned a great deal more about 
visual perception in the last 75 years.

I'm thinking that the basis for modern color management is coming from 
the International Color Consortium (ICC) and the ISO:

http://www.color.org/icc_specs2.html
> It's no coincidence that computer graphics standards have striven for and now essentially settled on a 24 bit colour space. It is because it is possible to represent the CIE (or all other) color spaces within the color cube defined by 256x256x256 values 
I have to disagree on this. Unfortunately I was there (jeez I'm old) 
working in the field when some of these choices were being made. Back 
then (late 1970s to early 1080s) we were still having the octal vs. 
hexadecimal debate (remember the DEC PDP-8?). The reason that computer 
graphics are eight bit is just money and technology. Eight bits was as 
wide as we could go at the time. Sixteen bits wasn't available and 
therefore wasn't even an option. Back then 8K of memory cost more than a 
car. Remember back to 1987 when IBM released the VGA monitor 
specification. Sixteen or 256 colors out of 262144 (six bits per 
channel). Took special chips which were worth making because the extra 
two bits cost a fortune.

We had eight bit color for a long time because many in the industry 
thought it sufficient. But this predates programs like Photoshop. When 
people started using computer photo editors they began a steady push 
toward 16 bits per channel. The fact that memory and disk space today is 
for all intents and purposes free just rids us of a possible constraint.
> and, in a sense, this color space represents a model of everything that we can see. I know that this is in a sense a subjective statement since perception is in part a pyschological phenomenon - look at things like color constancy as an example - a camera will record the light density changes across a wall, while our eye/brain system will often simply see a white (or other colour) surface (unless we know about colour constancy and look again). 
Color constancy is interesting. Since 1931, we've discovered a bit about 
how it works. The brain apparently contains some black point 
compensation circuitry that is separate from the rest of the visual 
cortex. Really most of us see in RGBK. Without it, moving from shadows 
to direct sunlight would result in perceived color shifts that would 
make our heads spin. And it's one of the reasons that our visual system 
is so bad at quantifying individual colors.
> So the very establishment of an international colour standard was in one sense an attempt to quantify the human visual system which is an attempt that can only approach the visual system. So I regard our attempts to produce some sort of 'definitive' and all encompassing reproduction system to
> be doomed to failure - largely because what we think we see  depends a lot on contextual information and not some kind of application to a number of 'levels'.
>   
The problem with standards is that they are just that - standards. When 
you learn more it's difficult to break free of the now outdated standard 
because you've build up so much useful stuff on top of the standard.

> Please feel free to correct me if I'm going off at a tangent here but I feel that a lot of this talk about bit depths and image size is about a pursuit of a technical excellence that begins to depart from the actual process  of vision.
>   
I know some experts and realize that I'm not one. I'm not correcting 
anything. I'm just giving you my perspective which is based on my 
experience and my reading. I think we all want the same thing which is, 
in the end, better prints.
--
Bruce Watson