Color management myths and misinformation video

[digitaldog]

Oh, your triangle explanation is definitely in the "number of angels on the head of a pin" territory.
If "sticking with words like gamut, range, and volume" you are DEFINITELY weaseling out of using plain English. You are hiding behind another layer that needs explanation (gamut, range...). All this just to avoid saying simply that Adobe RGB does have more discernible, actual colors, while at at the same time having the same number of potential colors as sRGB.

Slobodan, the state of New Mexico is vastly larger than the state of New York. Yet the population of humans in New York is much, much larger than the population of humans in New Mexico. The size of a state doesn't correlate to the number of humans who live there. However, thanks to Ted Turner who lives in NM, the population of Buffalo there is much larger than the population of Buffalo in NY.

The distinction between the population of humans and buffalo is analogous to this discussion of color gamut and number of colors in this way: Buffalo could be the gamut volume, humans could be the number of colors. This is why when I described populations above, I was very clear about what I was talking about, humans or Buffalo! It isn't weaseling out of using plain English! It is important to be specific when possible or there is potentially confusion.

Is Gamut Volume the same as number of colors? I'm working on that. Are humans and buffalo's the same (no).
Here's my current undstanding. The number of colors of images, which is what we're concerned with, is an attribute of encoding. The gamut volume isn't the same. There is no question that Adobe RGB (1998) has a larger gamut volume and a larger gamut than sRGB. Does that mean it has more colors? I'm not so sure. I don't think so but based on some tests I did with ColorThink, which could be calculating 'unique colors' incorrectly, I'm not ready to say one way or the other.

If someone says "Adobe RGB (1998) has more colors than sRGB", that may not be any more accurate than saying "New Mexico has a larger population than New York" when the assumption is we are talking about humans and not buffalo. And that's important. Otherwise we take the statement incorrectly and as fact.

[Eyeball]

If "sticking with words like gamut, range, and volume" you are DEFINITELY weaseling out of using plain English. You are hiding behind another layer that needs explanation (gamut, range...). All this just to avoid saying simply that Adobe RGB does have more discernible, actual colors, while at at the same time having the same number of potential colors as sRGB.

It can take a certain amount of work to come up with an explanation that is both easy-to-understand and reasonably correct, but I think it can be done.  "Range", "scope", and "volume" all seem to me to be pretty common, easily understood words that correctly convey the idea if "gamut" is too strange.

Also Slobodan, if you're subtly trying to imply with all this dialog that "that guy's video" was in layman's terms territory and we should have given him a break, I'll remind you of the following:

- the "number of colors" was NOT the only error.  He expressed very clearly that AdobeRGB and sRGB have the same boundaries and the only difference is the number of colors within those boundaries.  It's just wrong and misleading.

- the whole rainbow thing was completely unnecessary in the first place.  All he needed to say was "I recommend you stick with sRGB because AdobeRGB can get converted incorrectly by you, your printing company, or your web audience resulting in a dull image.  You can re-consider using a larger color space like AdobeRGB as you learn more about color management and are better able to reduce the risks while better appreciating the benefits."  He basically did what he accuses all of us doing by trying to impress people with charts and terminology for no good reason.

[Slobodan Blagojevic]

Buffalo city or buffalo animal?

[digitaldog]

This is basically what Chromix is doing when it uses a Delta-e value to define the area of a "color".  Just be aware though that this is placing a human perceptual limitation on what a color is.

But that's pretty important. As already pointed out, color, is a perceptual property. So if you can't see it it's not a color. A coordinate in a "colorspace" outside the spectrum locus is not a  
color. Color is not a particular wavelength of light. It is a cognitive perception, the excitation of photoreceptors followed by retinal processing and ending in the our visual cortex, within our brains. As such, colors are defined based on perceptual experiments. And from that, we get deltaE.

[digitaldog]

Buffalo city or buffalo animal?

Animal.

[fdisilvestro]

All possible colors in sRGB can be described in Adobe RGB
Some colors in Adobe RGB are out of sRGB

Result: More colors can be described in Adobe RGB

Please don't be fooled by the numbers, if using 8-bit or 16 bit or whatever. I still remember how long it took even for the "authorities" to understand that ETTR was about Signal to Noise Ratio (SNR) and not about numbers

[digitaldog]

Also Slobodan, if you're subtly trying to imply with all this dialog that "that guy's video" was in layman's terms territory and we should have given him a break, I'll remind you of the following:

I don't know why you have to and why Slobodan appears to be deliberately obtuse on the subject (sorry).   
Slobodan, you stated: If you look at the triangles, how can you NOT say that the bigger triangle (Adobe RGB) has more colors? I gave you an analogy with the size of a state and the incorrect assumption that it therefore has more people which isn't true. Your reply was to ask if I was referring to a city or an animal. Do you really want to understand this and add to the conversation or just always play (and not always well) devils advocate. The triangles are of differing size just like the two states. Assuming that means the larger triangle has more colors is faith based on your part, unless you can prove otherwise.
If you look at the US states, how can you NOT say that the bigger state (New Mexico) has more people than NY? Because it isn't true!

[Slobodan Blagojevic]

... Also Slobodan, if you're subtly trying to imply with all this dialog that "that guy's video" was in layman's terms territory and we should have given him a break...

Wasn't my intention initially, but now that you mentioned it...   No, seriously, I am trying to understand it myself, and find an explanation in plain English... it might be just a coincidence that we come up with a similar phrase as the-one-we-promissed-we-won't-mention-his-name.

...All he needed to say was "I recommend you stick with sRGB because AdobeRGB can get converted incorrectly by you, your printing company, or your web audience resulting in a dull image.  You can re-consider using a larger color space like AdobeRGB as you learn more about color management and are better able to reduce the risks while better appreciating the benefits."  He basically did what he accuses all of us doing by trying to impress people with charts and terminology for no good reason.

I like that... however, you had the benefit of revising the first, crude attempt, seeing his mistakes first. In other words, it is always easier to edit a negative example that to come up with it first. In general, of course, not that I doubt you are a guy who could have come up with a similarly elegant explanation on his own.

[Slobodan Blagojevic]

... Your reply was to ask if I was referring to a city or an animal...

Because you wrote Buffalo NY with a capital B? And there is such a city and it is in the state of NY. I was just clarifying for the purpose of analogy. No games here.

... Do you really want to understand this and add to the conversation or just always play (and not always well) devils advocate....

I do occasionally do that, though not this time. As I said to Eyeball, I am trying to understand.

...The triangles are of differing size just like the two states. Assuming that means the larger triangle has more colors is faith based on your part, unless you can prove otherwise. If you look at the US states, how can you NOT say that the bigger state (New Mexico) has more people than NY? Because it isn't true!

Ok... let's stick to this analogy. I do get it that two states might have different area but same (and definitely different) population. But if they have the same population, doesn't it mean that population density is different, i.e., there is more space between humans in the larger state? And yes, in that case, I am driving at the statement: "more space between pixels," but not to play games.

[mrenters]

I don't know why you have to and why Slobodan appears to be deliberately obtuse on the subject (sorry).   
Slobodan, you stated: If you look at the triangles, how can you NOT say that the bigger triangle (Adobe RGB) has more colors? I gave you an analogy with the size of a state and the incorrect assumption that it therefore has more people which isn't true. Your reply was to ask if I was referring to a city or an animal. Do you really want to understand this and add to the conversation or just always play (and not always well) devils advocate. The triangles are of differing size just like the two states. Assuming that means the larger triangle has more colors is faith based on your part, unless you can prove otherwise.
If you look at the US states, how can you NOT say that the bigger state (New Mexico) has more people than NY? Because it isn't true!

Andrew,

Are you saying that the colour points aren't uniformly distributed throughout the gamut volume?  I can understand that the population of a state may not be uniformly distributed, but surely points within a gamut are.

Martin

[digitaldog]

Because you wrote Buffalo NY with a capital B? And there is such a city and it is in the state of NY. I was just clarifying for the purpose of analogy. No games here.

OK. Sorry for the confusion.

Ok... let's stick to this analogy. I do get it that two states might have different area but same (and definitely different) population. But if they have the same population, doesn't it mean that population density is different, i.e., there is more space between humans in the larger state?

I don't think so, not necessarily. Take NM. There is just a tad over 2 million in that very large sized state. The vast majority of the population reside in two cities (Albuquerque and Las Cruces) and there are huge areas with no population. Not sure what that has to do with color gamut. The space between colors in an image in a color space would again be an attribute of the encoding.

[fdisilvestro]

Andrew,

Are you saying that the colour points aren't uniformly distributed throughout the gamut volume?  I can understand that the population of a state may not be uniformly distributed, but surely points within a gamut are.

Martin

Gamut maps are not necessarily perceptually uniform, which is equivalent to say that identifiable colors are not uniformly distributed in a gamut. In any case, any plot 2D or 3D of sRGB and Adobe RGB will show that sRGB is a subset of Adobe RGB, so I don't even understand how this can be a discussion

[digitaldog]

Are you saying that the colour points aren't uniformly distributed throughout the gamut volume? 

No, not at all. At least I don't believe that is the case. I'm saying the size of the triangle doesn't have anything to do with the number of colors. Slobodan and others see a larger gamut and assume a larger number of colors.
36 inches isn't bigger than a 1 yard simply because 36 is a larger number than 1. Or to put it another way, as I tried to explain to Gary on his video site, associating a larger gamut with more colors is like suggesting a gallon of water has more colors than a quart of water. The size of the container and the contents don't mean there are more colors.

[mrenters]

Gamut maps are not necessarily perceptually uniform, which is equivalent to say that identifiable colors are not uniformly distributed in a gamut. In any case, any plot 2D or 3D of sRGB and Adobe RGB will show that sRGB is a subset of Adobe RGB, so I don't even understand how this can be a discussion

That's what's confusing me.  If we have an sRGB volume and we look at the 16.7 million colour points in it, I think we all agree we can only perceive a subset of those because some of them are just too close to see any difference.  If we look at an AdobeRGB volume and look at the 16.7 million possible colour points in it, it would seem logical to me that since they are spaced further apart because they need to cover a larger volume, that we should be able to perceive a greater number of them.  Is this not the case?

Martin

[fdisilvestro]

That's what's confusing me.  If we have an sRGB volume and we look at the 16.7 million colour points in it, I think we all agree we can only perceive a subset of those because some of them are just too close to see any difference.  If we look at an AdobeRGB volume and look at the 16.7 million possible colour points in it, it would seem logical to me that since they are spaced further apart because they need to cover a larger volume, that we should be able to perceive a greater number of them.  Is this not the case?

Martin

It is the case, you will be able to perceive a greater number of them. My point is that when you look a graph or a plot of the color spaces, you cannot conclude by the relative difference in size the amount of additional colors in the larger space. E.G a 2D plot shows one gamut being 20% larger than the smaller gamut -> There are not necessarily 20% more identifiable colors, some more colors for sure.

[GWGill]

Adobe RGB (1998) has more colors than sRGB.

Comments?

This seems to be more about introducing the concept of a perceptually uniform colorspace, and the distinction between such a space and non-perceptually uniform spaces. Once perceptually uniform space is understood, then a gamut volume measured in such a space makes sense.

[digitaldog]

If we have an sRGB volume and we look at the 16.7 million colour points in it, I think we all agree we can only perceive a subset of those because some of them are just too close to see any difference.  If we look at an AdobeRGB volume and look at the 16.7 million possible colour points in it, it would seem logical to me that since they are spaced further apart because they need to cover a larger volume, that we should be able to perceive a greater number of them.  Is this not the case?

That seems reasonable but...

It is the case, you will be able to perceive a greater number of them. My point is that when you look a graph or a plot of the color spaces, you cannot conclude by the relative difference in size the amount of additional colors in the larger space. E.G a 2D plot shows one gamut being 20% larger than the smaller gamut -> There are not necessarily 20% more identifiable colors, some more colors for sure.

... And of course, the image we render into those color spaces have to play a role. A Granger Rainbow is one thing, something from a digital capture?

It sure seems somewhat dangerous to say "Adobe RGB has more colors than sRGB" if I've gained anything from this very enlightening thread. Saying the Gamut Volume is larger seems safe. 
I still think Mark's post is ultra important:

AdobeRGB and sRGB are just spaces, they don't inherently have any information (other than specifications for primaries, white point, etc). Until you actually have a pixel, there isn't any information.

[digitaldog]

This seems to be more about introducing the concept of a perceptually uniform colorspace, and the distinction between such a space and non-perceptually uniform spaces. Once perceptually uniform space is understood, then a gamut volume measured in such a space makes sense.

That sounds very reasonable. There is some debate if Lab is truly a perceptually uniform colorspace.

[Slobodan Blagojevic]

That's what's confusing me.  If we have an sRGB volume and we look at the 16.7 million colour points in it, I think we all agree we can only perceive a subset of those because some of them are just too close to see any difference.  If we look at an AdobeRGB volume and look at the 16.7 million possible colour points in it, it would seem logical to me that since they are spaced further apart because they need to cover a larger volume, that we should be able to perceive a greater number of them.  Is this not the case?

Martin

This is the most reasonable and comprehendible explanation so far for a layman like me. However, if true, it still means, in plain English, that there are more colors in Adobe RGB than in sRGB (that we can see).

[fdisilvestro]

something from a digital capture?

Take a photo of a standard 24-patches Color Checker
sRGB = 23 patches accurate, 1 cyan patch not accurate (the original color is out of sRGB)
Adobe RGB = all 24 patches accurate

24>23 => for a color checker image, there are more colors in Adobe RGB than sRGB

Other examples of everyday life colors outside of sRGB:

School bus yellow -> it is outside sRGB, maybe even Adobe RGB

Rainbow: as I understand, a rainbow may be represented as the boundary of the horseshoe (exept the straight line of "puples") in the cromaticity diagram, so neither sRGB nor Adobe RGB can represent "accurately" a rainbow.