@bitsgalore from my (even more limited) understanding, JPEG compression relies on downsampling the chrominance components, so I thought all JPEG files were transformed into YCC(K)... Would the compression actually behave differently depending on the input color space?