04-27-2014, 02:46 PM
Dem cats, and dat colour theory.
I will try and answer your questions with how I understand colour theory (Yes I did an image on colour theory using no colours, so what? xP)
Basically, the 'tools' in photoshop, really only show what is happening with digital colour algorithms. They show only a bare approximation of what happens in real life. For example, if you want your picture to stay true to your gray scale, you should convert the image to grayscale in the mode tab, not desaturate it with the ctrl+u slider or a saturation layer, these all work for an approximation, but they arent 'true' or as true as converting it to greyscale. If you want to test this out you can take a colour image, convert one to greyscale, and desaturate the other, and see the difference, then try colouring them with colour layers. I think it works for 80% of colours, but there are some hues and values it messes up. I haven't done this experiment in ages, so I wouldn't be able to stand behind this 100%, I know it messes some stuff up, but not entirely sure what anymore lol.
Anyway, that was a tangent. Basically, because of how screens work and programs work, I think the computer has a lot more information to play with in the lower values, rather than the lighter values where it just seems to fade so shades of white, where as in the darker values, you can get a myriad of different colours. The colour gamut of screens is also something to consider when talking about this. For example, an IPS apple screen will show a lot more information in the blacks, than say for example a standard cheap chinese screen that comes with most computers. This is why calibration is so important.
I think it is also because of what 'saturation' is physically. Colour works by reflecting light, and a colour is when one part of the spectrum is absorbed and one part is reflected back. The purity of this is basically saturation (the simplified graphs on the left which shows what is reflected back or in the case of screens 'outputted').
Therefore, if this difference can only be between 0 and 1 (in terms of value, 0 being black and 1 being white) then the more colour that gets reflected, across the entire spectrum, means that there is less hi-end differences available. If each globe was already outputting 80% if it total output, which is about a .8 on a value scale (to test this put in 200 in each of the RGB fields in the colour picker) then the max difference you can have at this level is 20%, one or two globes can be at 100% output, but the other is at 80%, so your max saturation is one fifth of what you had to play with when a globe was at 0 for example. (This isn't entirely true, because if two were at 100 and one at 80 your value would be higher, around a .9, but for arguments sake I think this is simpler)
That explains why you cant have more saturation in the upper part. Now the bottom part.
A completely saturated blue (one globe at 100%) is a value of around .3, a green globe is around .9 and a red is about .5
This means that if you are making a colour with a value of .3 or less, so a dark colour you can use the entire strength of the blue globe, a third of the green and more than half of the red (again, this is simplified because of the additive effect of the coloured lights, but I couldnt be bothered doing the maths to work out the true value which would probably be around a .6)
So for a .3 blue, you can still reach total saturation, you can get less saturation out of a colour that require green, since you only have a third to play with until you go brighter than .3 but if you are combining reds and blues, most of their spectrum exists down there. But in order to get a bright blue, you have to go into the cyans, which are closer to green, because the blue cant give you the brightness you need on your own.
I think this wall of text makes sense, its mostly been stream of consciousness, and I don't think most of it matches up with what I originally drew in my diagram... But thats how colour works in my head... Not sure if it is right, but I haven't disproven myself yet. Someone more experienced and smarter than me will probably say its all wrong.
I will try and answer your questions with how I understand colour theory (Yes I did an image on colour theory using no colours, so what? xP)
Basically, the 'tools' in photoshop, really only show what is happening with digital colour algorithms. They show only a bare approximation of what happens in real life. For example, if you want your picture to stay true to your gray scale, you should convert the image to grayscale in the mode tab, not desaturate it with the ctrl+u slider or a saturation layer, these all work for an approximation, but they arent 'true' or as true as converting it to greyscale. If you want to test this out you can take a colour image, convert one to greyscale, and desaturate the other, and see the difference, then try colouring them with colour layers. I think it works for 80% of colours, but there are some hues and values it messes up. I haven't done this experiment in ages, so I wouldn't be able to stand behind this 100%, I know it messes some stuff up, but not entirely sure what anymore lol.
Anyway, that was a tangent. Basically, because of how screens work and programs work, I think the computer has a lot more information to play with in the lower values, rather than the lighter values where it just seems to fade so shades of white, where as in the darker values, you can get a myriad of different colours. The colour gamut of screens is also something to consider when talking about this. For example, an IPS apple screen will show a lot more information in the blacks, than say for example a standard cheap chinese screen that comes with most computers. This is why calibration is so important.
I think it is also because of what 'saturation' is physically. Colour works by reflecting light, and a colour is when one part of the spectrum is absorbed and one part is reflected back. The purity of this is basically saturation (the simplified graphs on the left which shows what is reflected back or in the case of screens 'outputted').
Therefore, if this difference can only be between 0 and 1 (in terms of value, 0 being black and 1 being white) then the more colour that gets reflected, across the entire spectrum, means that there is less hi-end differences available. If each globe was already outputting 80% if it total output, which is about a .8 on a value scale (to test this put in 200 in each of the RGB fields in the colour picker) then the max difference you can have at this level is 20%, one or two globes can be at 100% output, but the other is at 80%, so your max saturation is one fifth of what you had to play with when a globe was at 0 for example. (This isn't entirely true, because if two were at 100 and one at 80 your value would be higher, around a .9, but for arguments sake I think this is simpler)
That explains why you cant have more saturation in the upper part. Now the bottom part.
A completely saturated blue (one globe at 100%) is a value of around .3, a green globe is around .9 and a red is about .5
This means that if you are making a colour with a value of .3 or less, so a dark colour you can use the entire strength of the blue globe, a third of the green and more than half of the red (again, this is simplified because of the additive effect of the coloured lights, but I couldnt be bothered doing the maths to work out the true value which would probably be around a .6)
So for a .3 blue, you can still reach total saturation, you can get less saturation out of a colour that require green, since you only have a third to play with until you go brighter than .3 but if you are combining reds and blues, most of their spectrum exists down there. But in order to get a bright blue, you have to go into the cyans, which are closer to green, because the blue cant give you the brightness you need on your own.
I think this wall of text makes sense, its mostly been stream of consciousness, and I don't think most of it matches up with what I originally drew in my diagram... But thats how colour works in my head... Not sure if it is right, but I haven't disproven myself yet. Someone more experienced and smarter than me will probably say its all wrong.