Profile Color Gamut Question

[OraDBA]

I apologize in advance if I don't describe this well, but here goes: all other things being equal (e.g., ink quality, longevity, etc.), will a printer profile with a larger/wider color gamut create a better print than one with a smaller gamut? I use Profile Prism to create custom printer profiles and I'm wondering whether it's a valid comparison to take the gamuts created by two different 3rd Party Ink maker profiles and see which has the wider gamut. Will (or should) a profile with a larger/wider gamut result in a print with better dynamic ranges, less clipping, etc.?

TIA,
Dylan

 

[Grandad35]

Dylan,

There are other forums (e.g. dpreview) where the author of Profile Prism (Mike Chaney) sometimes answers questions such as this. Even if he does not, there are lots of other people on that forum who are very knowledgeable in this area.

Let us know what you find out.

 

[OraDBA]

Grandad,

Funny, I was hoping you'd jump right in here and supply the answer! =8^) Afterall, I used one of your other posts that referred me to that http://www.iccview.de site. Some wierd profiling behavior got me to thinking there might be some substance to this. I will put a similar post on dpreview and see if I get any bites.

Cheers,
Dylan

 

[Grandad35]

Dylan,

I don't mind answering, but there are experts on the other forum who know a lot more than me about the details of profiling software.

Load the test image that you print when generating a new profile into PhotoShop and use the color picker to display the color values of the various color swatches. You will find that this image contains values at every corner of the R/G/B color cube from 0,0,0 to 255,0,0 to 0,255,0 to 0,0,255 to 0,255,255 to 255,0,255 to 255,255,0 to 255,255,255 and everything in between. When you print these test charts, these raw color values are sent to the printer to see what colors are actually printed for each of the target colors.

It is common knowledge that even today's best printers cannot print every color that is contained on these test charts, and that there are limitations in the printer's range of colors that can be printed, called the "color gamut". One of the functions of profiling software is to determine the limits of colors that were actually printed (the printer's gamut for this printer/ink/paper), and to include the color gamut information in the profile. Another important function is to determine how the colors should be handled as they approach or exceed the limits of the printer's capabilities - this information is generated for different "rendering intents" that give the profile a hint how you want these out of range colors to be handled. Finally, the profile must define a "look up table" that basically says "If you want color 'A' on the finished print, then you have to send these color values to the printer to actually get that color".

With that as a background, what happens when you use the wrong profile? Your driver will send data to your printer as though it were the printer for which the profile was developed instead of your printer. Unless the two printers are very similar (in which case the gamuts would also be similar), you will not get accurate colors. Since the purpose of a profile is to give accurate colors, this isn't a good thing. Can you drive a SUV as though it's a sports car? Yes, but only if you want to see the world by looking upside down. If there were such an easy, magic way to extend the color range of a printer, the printer suppliers would probably have tried it by now and made it a standard feature.

Do a search on the other forum for scanner based profiling systems, and you will find that some people love them, but that other people have troubles getting good profiles.

 

[OraDBA]

Grandad,

Thx for the info. Perhaps I'll share more details on why I became interested in this...You often see people with Canon printers who claim that once they go to non-OEM inks, they start getting really bad red/magenta color casts and/or oversaturation in their prints. You know, like really red skin tones, etc. I tried tweeking them, but the more I shifted things one way, others got out of whack.

So I started thinking that maybe one of the reasons my colors were off was because the batch of 8.5x11" paper that I was using to make my profile target was a slightly different formulation than the 4x6" paper I often printed to. So per the Profile Prism (PP) instrux, I printed the TIF target on 4x6 and I scanned at a higher resolution (it suggested 400dpi, but I used 600dpi instead). When the process was complete, I printed using the new profile and nearly fell out of my chair: the oversaturated red skin tones were GONE and things looked great! How could this be?

Mike is still looking into it, but in the meantime, I decided to compare the color gamuts of the original profile I'd made that was producing the oversaturation to the new profile I'd made using the smaller target at a higher scan resolution. What I saw was that the new profile's color gamut extended much further "south" (technical term!) into the blues & grays. I think this is particularly significant for me since I tend to shoot a lot of things in the shade which results in "cooler" WB settings. Then again, this assumes that I have even a small clue about that which I speak and that this makes any sense at all (neither are assured)...I'm not sure how to post images on this forum or I'd put a graphic up to show what I'm talking about.

So all this got me to thinking about what the impact of using different inks might be and whether an ink/paper combination that produced a wider color gamut than another might be "better."


Thx again,
Dylan

 

[Grandad35]

Dylan,

Here is a link on how to post images on this forum:
http://www.nifty-stuff.com/forum/viewtopic.php?id=36

If you are already talking to Mike Chaney about this, you are talking to the right person.

If you are using Kirkland 4x6 and Kirkland 8.5x11 paper, you should be aware that the 8.5x11 apparently has an Optical Brightening Agent (OBA) that converts the UV in fluorescent light and and sunlight into a blue-white light (similar to what happens to some white objects inder a "black light"). This makes the paper "whiter" when UV reaches the paper. The 4x6 paper does not have an OBA (it does not fluoresce under a black light), and it appears warmer even though a scan under the spectro shows that it should be cooler. Just a note - the OBA can "fade" fairly quickly when the paper is exposed to UV, so it may not give the same benefit over time. Also, any UV protective coating applied to your prints will prevent UV from reaching the OBA, making the paper appear warmer.

If a given ink/paper combination results in a wider gamut for a given printer, then that is a good thing - I misunderstood what you were doing.

 

[OraDBA]

Grandad,

Ironically, the paper that was casting all the reds was the Canon Photo Paper Plus Glossy. With all OEM ink, the colors are spot-on. But with my new Hobbicolor ink, there were oversaturated red shades particularly in the skin tones and they were driving me bonkers.

If this upload works, there should be one of those ICC gamut comparisons from iccview. The one in color is the one that produces the correct colors/skin tones. The other one (that doesn't extend "south") is the gamut of the profile that produced all of the horribly oversaturated skin tones.

Cheers,
Dylan

 

[Grandad35]

Dylan.

I would be interested in hearing what Mike said about this gamut map.

It appears that an extreme color value was measured for at least one of the colors on the test chart, and that the profiling program accepted it/them. Notice how the "southern extension" consists of perfectly straight lines that are "conically parallel" (if there is such a description). It is very difficult to believe that a printer's gamut could actually behave like this.

In the final analysis, a profile is supposed to give prints that reflect what you see on a calibrated monitor. IMHO if it does that it is a good profile, regardless of what the gamut map looks like.