CHROMiX

Profiling the Human Eye

As a first post, I hope I’m not ‘abusing’ the forum by bringing up a non-hardware topic. The eye is, after all, a scanner :wink:

The whole issue of front-to-back system profiling seems to have (so far) ignored the user/viewer in the equation. Granted, the ‘analogue’ front-end - the human eye - may not fit the normal quest for accurate colour, but, given that professional photographers should take into acount it is not only their eyes (and hence their subjective view of ‘accurate’ colour) that matters, but their customer’s, should there not be a means of eliminating the photographer’s/imageer’s biological non-linearity of his/her vision? Some form of eye-testing that could produce a ‘photographer profile’ and thus be included in the profiling system to allow for any one photographer’s eyesight/colour deficiencies? For example, an ‘viewer profile’ to combine with the monitor profile so that what the (potentially) colour-deficient photographer sees on screen will not affect target accuracies.

As ‘Colour’ is produced in the mind, not in the computer, I’d be surprised if different users do not have different ‘reference points’. Or does the human system have a self-calibrating mechanism which rules out such errors? (other than the fallible ‘auto-white-point adjustment’). Does constant viewing of ‘known’ colour values provide the mind with those accurate references, giving us all (eye problems aside) a near-perfect internal colour mechanism?

Recognising that colour management is a ‘best effort’ rather than ‘perfect system’, perhaps an eyesight profile would be of minimal practical use, but it would be interesting to hear other’s thoughts on the matter.

Mathew

P.S. Enjoyed the book Mr Rodney! :smiley:

I love this topic because it’s really the wrench in the gears of any colour management system: no matter how much CM you have, a customer could just say he doesn’t like it; and perhaps he really does see it differently than you. The PCS (profile connection space) does factor in the observer; however, it is an aggregated function of multiple human observers. This is called the Standard Observer, and it’s used to calculate Lab and XYZ values.

I guess it is, in principle, possible to generate a Custom Observer function and use that to calculate your PCS values. One could generate a Custom Observer function of her customer and build profiles specifically for that person. Your right though that this is probably impractical since the Standard Observer function does quite a good job, and even if the colours were absolutely accurate to that person’s eye, he could still say he doesn’t like it.