I’m trying to work out a method to create a multi-channel CMYK look where multiplying the four layers yeilds the source image and haven’t landed on a good method yet. If someone has one, I’d love to hear it.
I do not need this to be a “technically correct, converted in Photoshop and all gamut set for print” thing, just four “channels” that will multiply back into the original image.
You could probably extract C M Y K using the button at the button of Color Corrector node. What you do after that, I don’t know.
I think the issue may be that CYMK is subtractive.
What I’ve currently got going, that works fairly well (it’s not a perfect match to my extracted CMYK channels) is using the Red channel as the Cyan alpha, the Blue as the Yellow alpha, and the Green as the Magenta alpha.
Those multiply together to make a negative that gets negated back to an image that looks like I want. (which now that I think about it is kind of obvious, but anyway, im on a journey)
I went a step further and used all the suppress buttons on the color correct, then subtracted that from the source image to get a pure chroma, no luma image and split that out into the CMYK, and used the suppressed image as my K channel.
Here’s a blast from the past. I think it was shared on fxguide or flame-news. This is a batch setup to replicate the Technicolor process of the days long gone.
It basically takes the negative of the image, separates it into RGB channels (using CC node, because we didn’t have the Separate node back then), blends block colors as maximum (cyan for the red, yellow for the blue…), combines them with multiply and negates the whole thing at the end.
You adjust the block colors to affect the outcome.
You can use a luma key to adjust your blacks but I left the setup as it is from back 2001, when it was created… Rename the file back to .tar by removing .zip, forum software doesn’t allow for obsolete data formats…
Technicolor.tar.zip (30 KB)
Found this: RGB / CMYK conversion in GLSL · GitHub
Would be not be a lot of work to wrap this up as a Flame matchbox…. Could do it sometime later today.
Will have to verify that the math is correct, but no reason to believe it isn’t.
I was thinking about his very setup as as I screwing around yesterday. Nice to see it still exists!
I’m intrigued by this concept. What inspired you to go down this path?
As print is subtractive, and unlike a calibrated monitor where you have a standardised light curve where you know what brightness the white point is, what benefit if there of changing to CMYK, which will effectively go through a transform to put the image back to RGB anyway before you can display it? Sure, the colour saturation will be close but the amount of light on the printed page could be so different depending on the lighting environment it is viewed in? I would have thought the colour science of going to print, transforming from RGB to CMYK after completing all the working viewing RGB on an RGB monitor, would have been as solid as completing a job in ACES? I’ve never thought about it bus is there a printing industry viewing environment standard that printers adhere to when viewing work?
To be clear, I’m not suggesting what you are doing or your thinking is incorrect, I am genuinely intrigued by your reasoning behind it. The why?!
CM Why? K
this tutorial by Texture Labs:
And yeah, it’s just for the look. A misaligned RGB image always looks like a misaligned TV, so I wanted to see if I could make an image feel more like a misaligned photo or screen print.
