Regarding an experiment of people viewing test prints, I think it would be an interesting exercise IF DONE CORRECTLY. There are several variables that would need to be addressed and the expertise and experience of the participants in photography and digital imaging would play a role. Here is what I would do and the results I would expect:
I would use a wide-gamut printing device, probably a nice inkjet, that has a gamut that exceeds sRGB significantly in at least some dimensions.
I would make all test prints with that device and using the same paper (probably glossy).
I would have some pairs of comparison prints side-by-side and other pairs separated by distance (other side or room, different room). Viewing conditions (lighting) should be as similar as possible.
I would have participants self-qualify as non-photographer, amateur photographer, or pro/serious enthusiast (not perfect but probably a reasonable compromise).
I would use various images for the test, more-or-less as I am indicating below. Which specific images are used is a pretty big deal.
Here is generally what I would expect:
- In general, the side-by-side comparisons would show the most difference. Comparison prints separated by space would register less consistent differences, especially as the difference between prints drops.
- Test #1 = Raw file developed as sRGB vs. that same image developed as AdobeRGB and then incorrectly converted to sRGB (assigning sRGB rather than converting to sRGB).
This is essentially what Gary was demonstrating in his video. In this case I would expect a high number of people, regardless of qualification, to see the difference and to identify the raw-to-sRGB image as the better image. Even comparison prints separated by distance should be detected as significantly different, assuming the test image has a mix of colors including some that are fairly saturated.
- Test #2 = Raw file developed as sRGB vs. raw file developed as AdobeRGB. Subjects in test image include a significant amount of color that is outside the sRGB gamut (specifically, greens, cyans, and yellows).
Here I would expect the number of people detecting differences to drop pretty significantly - particularly for non-photographers and for comparison prints separated by distance. The famous Gary Fong "dullness" is a minor factor here. It comes down to differences in color saturation of some colors, subtle hue shifts, and detail.
- Test #3 = Raw file developed as sRGB vs. raw file developed as AdobeRGB. Colors in test image fall completely within the sRGB gamut (a portrait, perhaps, of a person wearing neutral clothing).
In this case I would expect the number of people (correctly) detecting a difference would drop to a very small number - probably only pros/serious enthusiasts using the side-by-side comparisons. Here, the only difference should be very subtle differences in tonal gradations in gradients. Even those few participants identifying a difference may have difficulty choosing which one they consider "best".
Obviously, there are a lot of ways such a test can be "gamed".
Now as related to Gary's video:
- There is little or no relationship between what such an experiment would show and most of the criticisms being leveled at Gary's video. Most people are agreeing that sRGB is a pretty good choice for beginners and people who are not interested in or not ready for color management. Most people are criticizing Gary's video for how he explains WHY sRGB is a pretty good choice for those people. More than 4 minutes of the 5-minute video is not even necessary. If Gary wants to keep it simple, then keep it simple. He doesn't need to add a bunch of incorrect and confusing crap. If someone continues to ask "Why sRGB?", he can just say "trust me" and recite his resume. That's what he gets around to doing anyway. He might as well save 5 minutes and move on to the next subject.
- Gary has shown in the video comments and here that has at least one of the following:
A. Poor reading comprehension
B. Poor writing skills
C. Limited time resulting in A or B
D. A conscious intent of misrepresenting what others have written.
It appears likely to me that he will, at best, likely misinterpret any results of an experiment and, at worst, manipulate the test and/or results to suit himself.
- As evidenced by other videos on YouTube, he has been giving this same spiel for at least 3 years. He has not only painted himself into a corner on the subject, he has built a brick wall between him and the rest of the room while the paint dried. He will likely apply any trick necessary (denial, obfuscation, misdirection, insults, reliance on his resume, references to others making the same mistakes, etc.) to avoid admitting he made a mistake. Ultimately and unfortunately, he will likely use all this discussion as proof that his "simple" explanation is much better - that it is impossible for the "color nerds" to explain this in less than 100 pages. Of course this is wrong - the somewhat complex explanations are only necessary in trying to correct his "simple" mistakes. There are many, many alternative ways to explain this to a beginner audience in a simple AND correct manner that won't inhibit their more advanced learning in the future.