VRay 2.1 SP2 - feature/bugfix list?

[racoonart]

The environment fog is fine; your issue is solved trivially by increasing the fog subdivs to e.g. 32. I thought I had replied to that, my fault.

Yes, this helps of course. But for me it's still strange that different buckets can produce different luminance in the fog (with low subdivisions).

[pixela]

I wonder about the Environment Variable we created for rendering with the old DMC Sampler of Vray 1.5 to be able to get noise-free and RAM-free renders.
Should we still keep that or is the DMC Sampler in the update working fine?

[vlado]

I wonder about the Environment Variable we created for rendering with the old DMC Sampler of Vray 1.5 to be able to get noise-free and RAM-free renders.
Should we still keep that or is the DMC Sampler in the update working fine?

Comparisons between 2.20 and 1.5 SP6 produced the same noise levels in all of the tests I did, but it might still be useful to make a few tests of your own.

Best regards,
Vlado

[vlado]

Yes, this helps of course. But for me it's still strange that different buckets can produce different luminance in the fog (with low subdivisions).

I agree that it's somewhat unexpected, but it has its explanation. The problem is in the adaptive sampling of the fog. When the samples are too low, the adaptive algorithm may terminate too soon with a sligly incorrect result for buckets where there isn't a lot of detail. In buckets where there is some detail, more samples will be taken leading to a slightly different result.

Best regards,
Vlado

[racoonart]

I don't want to discuss about this too much in this thread (cause it's not the right topic ) but I'm really interested in understanding how vray works. If I understand you correctly there is a kind of "sampling level" per bucket. I thought Vray (and other engines) would need to compare pixels from bucket borders with their neighboring pixels from the next bucket - to eliminate issues like that. Otherwise there would always be some difference between buckets.

[vlado]

I thought Vray (and other engines) would need to compare pixels from bucket borders with their neighboring pixels from the next bucket - to eliminate issues like that. Otherwise there would always be some difference between buckets.

V-Ray does indeed compute an extra border around the buckets to ensure that there are no differences in the AA filtering. However each bucket along with its extra border it still calculated completely separately from the rest, they don't directly share any color information. (Sharing the information would create problems with multithreading and memory).

Bset regards,
Vlado

[PIXELBOX_SRO]

Dynamic memory works better now as per the fix list...

does this mean the memory consumption will be better now as per the issues found by Francesco Legrenzi some time ago?

[vlado]

does this mean the memory consumption will be better now as per the issues found by Francesco Legrenzi some time ago?

Yes, it should be better, at least as far as I could test.

Best regards,
Vlado