Min Shading Rate

[M.Max]

We will probably be removing most of those settings for future releases anyways... haven't seen anyone using the local sampling controls in a while.

Best regards,
Vlado

I wouldn't mind that if we will still have access to it in someway (maxscript, hidden rollout.. Etc).
Having the choice to tweak things manually can be helpful sometimes and I don't want V-Ray to lose its flexibility and become like that renderer (Corona I'm looking at you)

[AC5L4T3R]

I just opened an old file and someone had set the min shading rate to 4096, can anyone with a better understanding tell me why this is a good or a bad thing?

[joconnell]

This is a bad thing. Min shade rate is a forced value and it's driven by how many aa ray are shot at an object. If you have a really easy section like a flat surface, 1 aa ray might do it but you still get hit with 4096 samples. If you have a really fine edge where you're using something like aa 8, then 8 squared is 64 potential aa rays and then 4096 samples for each of those. Chaos chose 8 based on the average of all the scenes they profiled and felt it was the best balance of speed and quality, I can't think of many cases where 4096 would be appropriate!

[Karol.Osinski]

I just opened an old file and someone had set the min shading rate to 4096, can anyone with a better understanding tell me why this is a good or a bad thing?

I think that person mistaken MSR tab with the render resolution tab hah

[Neilg]

A while ago i was posting about how i'd done a few jobs where we kept the AA low and raised the min shading rate to very, very high values - it was relayed to me by others including lele and vlado that this was an unorthodox way to do things, and the minimum shading rate I was using was at insane levels that should never be used.
More importantly, the number I had which caused so much outrage was 128. I think 4096 is probably a bad mistake...

[Pixelcon]

I had to use something like 200 or 300 recently as I was using extremely large area style lights and the shadow falloff areas just would not clean up no matter what I did unless I increased msr value massively.

[^Lele^]

What was you max AA?
We suggest setting that higher, rather than MSR, as it will allow for *much* better adaptive behaviour from the sampler, resulting in lower render times to a given N.T.

[Pixelcon]

What was you max AA?
We suggest setting that higher, rather than MSR, as it will allow for *much* better adaptive behaviour from the sampler, resulting in lower render times to a given N.T.

50 or 100. It appeared to make no difference.

Actually I usually say I can't send files for analysis, but now I am able to. Would you be interested in a stripped version of this scene?

[^Lele^]

yes please, very much so!

[Pixelcon]

Ok. I'll get to it sometime this week. Which email should I forward it to?

[^Lele^]

The one in my signature should be fine.
I'll then share it with the relevant people (i.e. coders, support) should the need arise.

[Pixelcon]

I lied. I had pushed it up in an attempt to combat a strange behavior caused by setting VraySun's size to extreme values (above 100) and ended up fixing this by using a disc light to fake the sun.