Tweet
I'm working on a render that I am unable to show but it is a close up shot of a face (AlSurface) with displacement and hair (VRayHairMtl). There is very little direct light and most of it comes from GI.
My current settings are Min Subdivs: 2, Max Subdivs: 16, Threshold: 0.01, Min Shad. Rate: 9. This takes around 15hrs per frame. If I change Max Subdivs to 12, the render time increases to 18h. If I leave Max Subdivs at 16 and raise the threshold to 0.03 the render time decreases to 8h.
Raising the threshold makes a lot of sense, I'd expect more noise and lower render time. However I don't understand why lowering Max Subdivs increases the render time. Why would that be?
Here's my thinking, I hope it makes sense:
Is the reason because lowering max subdivs means less samples used for camera samples (because the min shading rate stayed the same). Thus it's harder for the renderer to get to the threshold of 0.01 and the renderer then uses up all 12 max subdivs to get to the threshold where with 16 subdivs the renderer might have stopped at 8 subdivs because there were more camera samples available?
My current settings are Min Subdivs: 2, Max Subdivs: 16, Threshold: 0.01, Min Shad. Rate: 9. This takes around 15hrs per frame. If I change Max Subdivs to 12, the render time increases to 18h. If I leave Max Subdivs at 16 and raise the threshold to 0.03 the render time decreases to 8h.
Raising the threshold makes a lot of sense, I'd expect more noise and lower render time. However I don't understand why lowering Max Subdivs increases the render time. Why would that be?
Here's my thinking, I hope it makes sense:
Is the reason because lowering max subdivs means less samples used for camera samples (because the min shading rate stayed the same). Thus it's harder for the renderer to get to the threshold of 0.01 and the renderer then uses up all 12 max subdivs to get to the threshold where with 16 subdivs the renderer might have stopped at 8 subdivs because there were more camera samples available?
