Test with Blue Noise Pattern
600 samples
PGI ON
Classic vs Blue Noise
Add two images here. Change ‘data-direction-horizontal’ to ‘data-direction-vertical’ for a vertical slider.
1 Like
barely any different, just a different noise pattern at sample count I feel
2 Likes
For me it all depends what it looks like in motion. For stills, I agree it doesn’t make much difference.
2 Likes
I fully expect it to make a bigger difference in other situations, yeah. In principle, it should be particularly performant in the context of denoising
but for example, blue noise does not work with noise threshold activated (only the first samples as mentioned in the PR), wouldn’t that mean more render time to achieve an equally clean render?
not necessarily. It’s gonna come down to a case by case test.
Anyone seen the artefacts with Light Tree?
Only work around seems to turn the feature off completely.
I can’t find a discussing about it.
Hey ! From what I understand, to get the most noticeable difference you should test with very low samples like 1, 3, or 10… At 600 all these samples are already averaged a lot and it’s indeed very similar. Since blue noise is about making the most uniform sample possible it’s effective only in the beginning so it’s mostly useful when you work in interactive mode, or when you’re using denoising or maybe with adaptive sampling…
I read that it’s not meant to work in interactive preview