It seems you are trying to do fairly steep color corrections in CMYK.
That is not a good choice.
As the image very likely was in some RGB-space originally it is more prudent to do the color manipulation in that space and then separate as needed for the various print processes.
Colorising can be done in a number of ways, some examples:
Solid Color Layers set to Blend Mode Color,
Gradient Map Layers, …
Edit: All naturally in many possible combinations, in particular with Curves Layers to address luminance where necessary.
Channel Mixer is certainly not an illegitimate choice but should, in my opinion, not be the first one for the task.
In any case:
You could use the same Channel for more than one other Channel and you could mix more than one original Channel into one resulting Channel.
Problems with editing CMYK images are that one can easily mess up the balance of different parts of the image and overshoot the allowed Total Area Coverage (TAC).
Thank you for taking your time. But I am doing the color manipulation in RGB. You think I am in CMYK because I wrote mixing the Red and the Blue is as mixing cyan and yellow. Because cyan and yellow makes green.
Manipulating color on things with detail it is not possible that the detail remains if you are completely switch the color. Only if you are using the channelmixer the detail stays. Afterwards if the color is in the ball park you can use HS or Selective color but before as I see it. So my question is on color theory which stays behind the channel mixer.
So my question is on color theory which stays behind the channel mixer.
The interface seems straightforward enough and as far as I can tell the values get added or subtracted according to the settings but are naturally limited to between 0 and 255.
Only if you are using the channelmixer the detail stays.
But please post an example (source image and target color) that you feel is impossible to achieve with anything but Channel Mixer.
But I am doing the color manipulation in RGB.
Sorry, I misinterpreted your original post.
Ok, I have to admit the advantage I expected by using the channel mixer is not existing. It is just a different workflow. There is a chance of destroying detail by using the HS but is not necessarily the case.
Otherwise by using the channel mixer for color manipulation and the selective Color for the endgame you are reducing the amount of wanging the sliders back and force. Therefor it seems to be a bit more reliable because it is more mathematic than controlling by eye and hand.
Photoshop often offers more than one, sometimes downright »many« ways to achieve a task and if you get the results you want nobody can call the method you choose »wrong«.
Me, I hardly ever use the Channel Mixer for recoloring but not only may our tastes differ but the requirements and source data for the work we do may be different, too.
Therefor it seems to be a bit more reliable because it is more mathematic than controlling by eye and hand.
I think it is worthwhile keeping in mind that humans do not have absolute vision and that different Color Spaces can carry different properties, so approaching color correction with a strictly mathematical approach may at times fall short of expected results.