![]() I thought I was doing exactly the same as before - I guess I wasn't.Īnyway, for the curious, here is a screen recording of what I saw earlier with in the color picker. The only way to get these shifts in the color picker are happening when using a too small histogram profile, e.g. Now I can't reproduce the cyan shifts any more. I tried one more time, with a release build instead, and not attaching the debugger. It seems to be the extreme lightness, that causes the hue shift. What's interesting to see though: the cyan shift is not actually a hue shift. I have no idea why the color picker behaves like that. I've thus set the output profile to linear Rec2020 RGB but can still reproduce. I know there are some more profile conversions when it comes to the histogram. I've also noticed cyan in the live color picker, just before the black clipping, but I am not sure if that's the same effect that happened to my previews. I haven't found a way to reproduce unfortunately. With the same settings, not all of them are showing the same casts. ![]() The small preview, the filmstrip and the main preview can still show a cyan image, although a much smaller area than before. But even then, I managed to get some strange effects. My first thought was that I din't build correctly, so I wiped build and config directories and tried again. I tested and could stills see some cyan shifts. Saturation pushed - filmic off (filmic on does not change the clipping) Linux, OpenCL active - but can reproduce without. Shouldn't increased chroma clipping result in white instead of black? Looking at the 2nd image, right corner, this would avoid the harsh transition.Why is there a hue shift to cyan? This doesn't botter me really, since it seems to not happen when filmic is switched on.I am not sure if my approach was doomed from the start due to testing with a png, but I can also reproduce with a tif file.ĮDIT: There are actually two things I am wondering: Pushing saturation skips the cyan conversion and just goes black.Īctivating filmic with default settings mitigates the cyan issue, since everything high luminance is desaturrated. Isn't that contrary to what's written in the manual? What surprises me, is that blue (hue 243) turns cyan (hue 185) before it eventually turns black when pushing chroma. I've downloaded one of the graphs from and pushed chroma respectively saturation (beyond reason, I know). I hope you don't mind to elaborate what's happening and whether this is the intended behaviour. This prevents the out-of-gamut colors that can quickly be produced when increasing chroma and saturation.īut to me, the color changes are rather surprising. ![]() For some learning exercise I've loaded some artificial test images and played with the color balance rgb module.Īt its output, color balance RGB checks that the graded colors fit inside the pipeline RGB color space (Rec 2020 by default) and applies a soft chroma clipping at constant luminance and hue. I am not sure if this is a bug, or if I am just too stupid to get my head around the color science. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |