The other thing I'm not clear on is the gamut of the device you are using to edit your picture. If the monitor you are viewing it on cannot display the range of colours in the colour space you are using, then whats the point? I think ProPhotoRGB must be far larger than your average monitor, I'm sure one can spend thousands on a higher spec monitor, but that is likely beyond the realms of the majority. I suppose my limited understanding of the topic is questioning why should we work in a colour space that we can't see the full extent of?

The other thing I'm not clear on is the gamut of the device you are using to edit your picture. If the monitor you are viewing it on cannot display the range of colours in the colour space you are using, then whats the point? I think ProPhotoRGB must be far larger than your average monitor, I'm sure one can spend thousands on a higher spec monitor, but that is likely beyond the realms of the majority. I suppose my limited understanding of the topic is questioning why should we work in a colour space that we can't see the full extent of?

So saying “yes, color is a continuous spectrum so the wider the range we can capture the better” is true IF we were working in the analog domain. However, we're working with digital, which is unable to sample infinite points but will now condense the continuous spectrum into only 16k or 4k or 256 discrete points.

So while srgb gives us a smaller range to work with, we are able to describe the colors in that range much more precisely, given the limited sample points we have.

As a simplified example, let's take a description of “blue” that ranges from say 0-100, continuously in real life. Say our sRGB can only cover the 35-65 range for this blue, while Adobe RGB can cover the 20-80 range. Adobe RGB can cover a superset of sRGB's colors, right? Not so in the digital domain.

Now for simplicity's sake, say we're working with 2-bit images, which gives us 4 possible values. With sRGB then, we could describe the points 35, 45, 55, 65. With Adobe RGB, the colors we could describe would be 20, 40, 60, 80.

Now imagine how the colors from subjects in the scene would map to digital values in either of these schemes. What if we had say a blue flower petal, whose tones ranged continuously from 35-45 (in our 0-100 real-life scale). With sRGB, this continuous spectrum would get flattened to one of two tones: everything from 35-40 gets mapped to 35, everything from 40-45 gets mapped to 45. What about Adobe RGB? Each discrete values takes big jumps within the spectrum, since it has to cover a much wider spectrum with the same number of points. So for this flower petal, its *entire* 35-45 spectrum gets mapped to the same value: 40. We get no detail for this flower, since everything maps to the same color.