Copy link to clipboard
Copied
Hi, I noticed SRGB, Adobe 1998 and Prophoto don't give the same results (especially in the darkest dark) when I modify my images with an exposure adjustment.
Is one of the color space where the results are more physically accurate?
Thanks
Copy link to clipboard
Copied
That's exactly what the Adobe Help states.
Copy link to clipboard
Copied
From what I understand about the exposure adjustment. In a 8bit or 16bit document, Photoshop takes the rgb gamma value from your current color space, convert them to linear, apply the exposure adjustment, then reconvert the linear values to the gamma color space you are working in. From what I understant most of the adjustment just apply the calculations directly to the gamma you are working in. But, I might have misunderstood.
Copy link to clipboard
Copied
Think of it this way. You could easily have an adjustment in ProPhotoRGB that would clip in sRGB or CMYK. There is a reason that Adobe warns you when placing a document with a different colorspace than your working document. I think the bottom line is that colorspace has an effect on how your images look.
Copy link to clipboard
Copied
Photoshop takes the rgb gamma value from your current color space, convert them to linear, apply the exposure adjustment, then reconvert the linear values to the gamma color space you are working in.
By @Guillaume5FC6
That's how I read those Help pages too. So in that sense, it should at least be consistent across color spaces.
But the problem is that these adjustments aren't properly defined to begin with. There's no indication of what it actually does to the RGB values, only a diffuse concept of "stops" that doesn't really apply here. A stop is doubling or halving the amount of light that hits a camera sensor (or film), and you can't translate that to a digital file. It doesn't apply. So you're left with "move the slider and see what it looks like".
Copy link to clipboard
Copied
Remember that most film and digital sensors have a non-linear response to light. There isn't a way to accurately model that because each camera and film stock is a bit different.