Skip to main content
Photography

Timeline for What are Color Profiles and where would I find information on using them properly?

Current License: CC BY-SA 3.0

14 events
when toggle format what by license comment
Dec 15, 2013 at 19:39 comment added Nothingtoseehere I never said Lab was unimportant, nor did I say to write it off. I did say that your assumption of how the process works was incorrect. If the color space is RGB, the transforms are RGB to Lab. Lab color has very little to do with the human visual system other that it was a failed attempt at perceptually uniform color space. Others have sprung up afterwards and WCS uses CIECam02 which was supposed to be an improvement, but alas it fails in many ways. So remember that in an RGB color space the bits needed to render colors properly are less than in Lab, it's about the bits not the space.
Dec 15, 2013 at 16:33 comment added jrista I am sorry to disagree, but Lab is critically important. You do understand that ALL color space transformations move through the Lab space, right? RGB color is first converted to XYZ, which is then translated into Lab, color space conversions occur there, then translated back to XYZ and converted back to RGB. You cannot write off Lab, nor the work that CIE did to design the Lab space so many decades ago. It is a critical part of the mathematical process of converting one color space to the other, which means that human visual perception is ALWAYS taken into account.
Dec 15, 2013 at 6:00 comment added Nothingtoseehere No one's writing off profiles, or saying that today's monitors are no good. I am suggesting that rending intents and gamut mapping must be taken into account when high quality color workflows are needed. I am also saying some of your answers and blog post reference's are inaccurate and incorrect. Hopefully the answer I posted will help clear up the holes, and serve to educate others that need accurate information on this somewhat technical subject.
Dec 15, 2013 at 3:27 history edited Nothingtoseehere CC BY-SA 3.0
add images to further clarify what ICC profile do. I think most people lean best if they see a picture So I hope these will help persons that still don't understand my answer.
Dec 15, 2013 at 1:18 comment added jrista However, Pro Photo RGB shouldn't be written off as too large, either. It doesn't matter if it is too large for screen or print...what matters is that it preserves the most natural Lab gamut coverage possible. ICM will deal with converting from Pro Photo RGB to whatever our output device is, and maintain accuracy for us...SO LONG as we haven't ruined that accuracy by first manually "downsampling" into something like sRGB first.
Dec 15, 2013 at 1:17 comment added jrista We can't tell the difference by looking at the image rendered on our screen...ICM deals with conversions between color spaces, such as Pro Photo RGB and the monitor's calibrated profile, automatically. We see the same color on screen regardless of the profile the image is tagged with. HOWEVER, if we PRINT the original Pro Photo RGB and the post-sRGB Pro Photo RGB images, the difference in that green color will very likely show up, as print gamuts tend to behave a little differently than colors rendered onto a screen. You ARE correct, sRGB is archaic, and probably shouldn't be used.
Dec 15, 2013 at 1:15 comment added jrista As for gamuts in general, it is a matter of extent, not total color count or anything like that. Pro Photo RGB is an extensive gamut. It covers the entire range of "visible light", the saturation extent, black point, and white points as derived by the CIE working group in the 1930s. If we take a pure 16-bit green, for example: 0, 65355, 0. In Pro Photo RGB that value maps to a naturally saturated green. If we convert to sRGB, that value now maps to a relatively more muted green. If we convert back to Pro Photo RGB, it still maps to a more muted green.
Dec 15, 2013 at 1:12 comment added jrista When it comes to rendering intents, things are not really all that cut and dry. In my experience (which is now greater than when I originally wrote my answer), you need to use the rendering intent that a printer ICC profile was designed for. In the case of my nice neutral white Red River Fine Art rag paper, I have to use Perceptual...its how the ICC profile was designed. Relative clips. However, Hahnemuhle, which provides a very similar paper, designes their ICC profiles for Relative Cholorimetric intent (no clipping). If I use perceptual, the color comes out wrong...not bad, just wrong.
Dec 15, 2013 at 1:10 comment added jrista Hmm, I am not sure all of this is correct. It doesn't take a $25,000 display to render a hell of a lot more than the AdobeRGB gamut. For example, Lacie had a $3000 30" RGB LED screen that could display over 130% of the AdobeRGB gamut, and it was only 200cd/m^2. There are professional grade 4k screens coming onto the market that theoretically support that much or more, again at well below 600cd/m^2 brightness. My Lumia 920 phone renders more than the AdobeRGB gamut, for Christ sake, and it was only $200! :P I also know that my aging 9500 II printer at a medium L* is better than AdobeRGB.
Dec 14, 2013 at 16:24 comment added Nothingtoseehere @mattdm Your right. I have edited my reply to provide a proper answer and clear up some of the unusual points in the answers above.
Dec 14, 2013 at 16:22 history edited Nothingtoseehere CC BY-SA 3.0
to more clearly answer the question and dispel some of the myths in the answers above mine.
Dec 14, 2013 at 15:25 comment added mattdm This answer has a lot of good information but it is kind of confusing because it doesn't directly answer the question but rather appears to refer to other answer and comments, but since Stack Exchange isn't a forum it isn't clear exactly which ones. Can you edit those sections to respond to the question directly, please?
Dec 14, 2013 at 15:15 review First posts
Dec 14, 2013 at 15:25
Dec 14, 2013 at 14:58 history answered Nothingtoseehere CC BY-SA 3.0