![]() | 1 | initial version |
View matrices were not significantly different. After running a few times the differences do indeed seem random.
The differences in daylighting that I got might be due to small changes in that randomness causing rays to intersect a low resolution BSDF on a significantly different "patch".