Understanding DMC calculation

Hi!

I’m following this tutorial on how to optimize V-Ray Image Sampler and DMC to reduce noise and mantein low render time:

If I understand correctly the VrayLight Subdivision is divided by the Image Sampler Max Subdivision. This is an example scene:
http://www.ossosso.com/scambio/shoot\_15.jpg
where 64/64=1 sample for VrayLight.

If I increase VrayLight Subdivision to 24 as expected the red areas became smaller like here:
http://www.ossosso.com/scambio/shoot\_16.jpg
again 576/64=9 samples for VrayLight

Here what I don’t understand: if I leave VrayLight Subdivision at 8 and increase Global Subdivs Mult. to 3, I should have always 24 subdivs for VrayLight, so the SampleRate result of this render should be the same as the previous:
http://www.ossosso.com/scambio/shoot\_17.jpg
But as you can see it isn’t; red areas are larger then the previous render.

I’d like to understand how “Global Subrivs Mult.” works because I have to manage complex scenes with a lot of materials and lights, and it’s quite time consuming to manage any single light subdivs and material reflection/refraction subdivs.

Thank you very much.

The global subdivs multiplier is not squared like the rest of the subdivs; you will need to set it to 9 to get the same result as the 24 light subdivs.

Best regards,
Vlado