Photo Seren wrote:
As far as I can tell it separates the image into RGB and CMYK chanels and then you can alter these values with the sliders to set end contrast.
Uh no...The color values are _NOT_ typical RGB/CMYK but specific values that more directly relate to real life. There is no Cyan, but there are Aquas (and no, they are not the same).
The Convert to Grayscale Option takes all 8 channels and tries to combine them in an optimal mix to maintain color contrast. Don't confuse this with tonal contrast. Often two strongly different colors will end up converting to B&W with very similar tones. ACR tries to maintain the apparent contrast even after grayscale conversion.
The individual sliders allow you to modify the mix of an individual color in the resulting conversion. Think of plus numbers as gain and minus as subtraction.
Ok... so where does it get these values from? And how do they affect the RGB pixel values of the image? Does the aqua slider for example affect equal values of the blue and green in the image?
The raw file holds information giving each pixel a red, green or blue value and then these are processed during demosaicing. So is it these values that are altered and the demosaicing algorithm affected by your choices in the grayscale mix panel or is the image demosaiced using pre-sets and then the values altered after...
Thanks for your help so far. I think I am starting to get my head round it!
Actually, in terms of the HSL/Grayscale panel, it's NOT working in an RGB space but in a Luminance/Saturation/Hue space where the color data is handled separately from luminance. I'm pretty sure it's working in CIE-XYZ but that's only a vague memory of what somebody said. The color Hue that is uses in the HSL/Grayscale panel are not spectral additive and subtractive primaries such as RGB/CMY. They are specific areas of the spectrum that more accurately represent real life photo scenes.
You should also prolly quite trying to think about this stuff as "demosiacing" the raw file. That actually happens under the hood in the background as part of the processing pipeline. The rest of the image manipulation ,ay or may not be done in a color space vs luminance separate from color. For example the Curves adjustment adjusts Luminance and Saturation but locks Hue. The sharpening is done only in Luminance. Luminance noise reduction is done in Luminance (well DOH) and Color Noise is done in color.
When the HSL/Grayscale option to convert to Grayscale is selected, the auto function tries to maintain optimal color contrast after the conversion. You can add or subtract the hue of the color by using the sliders...but what you are actually doing is adjusting the color under the hood, not the tonal curves.
Well that gives me a starting point at least, thank you! This is not really my area so much so it's hard to know where to start.
So are these colour spaces desided per image? what is the deciding factor about what areas are affected? Though it's not working in RGB it must have an affect on the final RGB values? It goes from RGB to whatever the other colour space is and back again when you save to tiff or other fornmat? or have I just further misunderstood?
Photo Seren wrote:
So are these colour spaces desided per image?
Camera Raw converts from the camera captured file into it's internal working space which is the chromaticity of ProPhoto RGB but with a linera gamma instead of PPRGB's native 1.8 gamma. But you really don't need/want to deal with that. When you are done and you process the image out into Photoshop, Camera Raw transforms the file through it's full processing pipeline and it comes out the back end in whatever color space you've speced in Camera Raw's workflow settings.
As far as what the HSL/Grayscale is doing to the RGB numbers, well one you convert to grayscale it really doesn't matter. This tool is really designed for visual adjustment and even if you could get the RGB numbers, they would be meaningless...