[argyllcms] Re: Question regarding gamut mapping for photographic images

Greg Sullivan schrieb:

Hi,
Referring to this extract from a post to newsgroup sci.engr.color by
tlianza, in thread "Chromacities of digital photo-exposer":

"The final reproduction gamma of a reproduced image from an sRGB image through 
a Frontier printer should have a relative gamma of 1.2 to 1.6

Hi Greg,

Argyll's perceptual intent involves a CIECAM02 appearance model. Depending on the source and destination viewing conditions that you specify, the relative luminance of mid-tones will indeed be changed by this transformation.

A "dim" surround in the src environment (typical CRT viewing conditions), combined with an "average" surround in the dst environment (typical for viewing reflection prints) actually result in a lightness mapping which corresponds to a relative gamma of circa 1.2 (but the other specified viewing conditions play a role as well), and if the surround of the src environment is assumed to be "dark" (e.g. projected slides), the lightness mapping of the CIECAM02 model will even correspond to a relative gamma of circa 1.5.

Anything less than that will yeild an image that appears flat and lifeless when 
viewed at interior lighting conditions.

IMO this statement rather refers to the image on the monitor. On a CRT in a "dim" surround, you need to display the image with darker mid-tones than on a print, being viewed in an "average" surround, in order that the printed image and the CRT image have the same appearance, i.e. in order that the CRT image does not appear "flat", compared to the print. And a projected slide (in a dark surround) needs to display the mid-tones even more darker, to obtain the same appearance.

and the extremes of sRGB should be mapped to the extremes of print material.

Depens. I would not be so sure. A perceptual intent is usually rather not expected to expand the source gamut to the destination gamut in those sectors, where the destination gamut is larger. I'm also not sure whether a mapping of the extreme points of the sRGB gamut to extreme points of the printer gamut might possibly introduce too large hue shifts, which are no longer pleasing. However, there are not fixed rules for gamut mapping - it's rather a matter of art, only the result counts, and "pleasing" is a matter of taste.

For instance, for my taste Argylls lightness mapping produces too dark mid-tones on devices which can't reproduce a very dark black, thus I've raised the lightness knee point. And I've also changed the chroma knee point to obtain less gamut compression for in-gamut colors.

Regards,
Gerhard




Other related posts: