color mapping with don't affect colors selected

[geminiPrevails]

Hi,

From what I understand: With a linear workflow, the choice of the color mapping type doesn't really have any effect on the colors or the tone of the rendered image (as the "dont affect colors" would be clicked). So would there be any reason to prefer one type of color mapping over another?

My guess is that these types would still be in effect in the sampling of the image. If that's the case would choosing any one of them over other types would yield a less noisy image in less or equivalent times?

Thanks

[Svetlozar Draganov]

You assumptions are correct - when don't affect colors option is activated selected color mapping is applied only for the sampling - color tones are not changed.
There are two benefits from that:

1.Export image in Linear Color Space
2.Better sampling (less noise) in dark areas of the image

In terms of speed/quality I've just run a few tests and I could say that each Color Mapping Type produce different render-time but in general they are very similar except for the Intensity Exponential Type which produce almost 3x longer render times. Here are how different types are arranged according to render-times.

1. HSV Exponential
2. Exponential
3. Gamma Correction / Intensity Gamma - same times, same quality
5. Linear / Reinhard - same times, same quality - this is actually expected cause by default Reinhard type works exactly as Linear type
7. Intensity Exponential

In terms of quality the chart will be exactly the same as the above but in reverse. This is due to the fact that better sampled image have better quality and longer render times.

[geminiPrevails]

Thanks Svetlozar for the detailed explanation! in the test I did linear seemed to yield less noise actually, but I doubt my scene was nearly representative as what you might be using to test.

One more thing I am wondering is how these parameters work when used in conjunction with clamping. For example I know that exponential color mapping would not let the extra bright regions to go overblown, does this mean that If I am to use exponential type along with clamping, less information is gonna get clamped - as they don't go overblown to begin with - resulting in conservation of more sampled values in the final render?

I am just trying to figure out if there is a 'better' choice to go with for exponential types, when using linear workflow.

Thanks!

[Svetlozar Draganov]

One more thing I am wondering is how these parameters work when used in conjunction with clamping. For example I know that exponential color mapping would not let the extra bright regions to go overblown, does this mean that If I am to use exponential type along with clamping, less information is gonna get clamped - as they don't go overblown to begin with - resulting in conservation of more sampled values in the final render?

Yes - this is correct. When exponential type is used instead of linear the final value for each pixels calculated with exponential type will be lower than the same pixel calculated with linear type therefore during the clamping stage less pixels will get clamped - more pixels will retain their original values.

[geminiPrevails]

I see, thanks a lot for the answers.