Dome Light - DMC Sampler VS Subdivision Sampler

[FSGFX]

Hi!

This is a thing that has kept me wondering for a while now, and it would be great to finally get an answer on this issue.

Has anyone else noticed, that when you light your scene with a simple dome light (HDRI or no HDRI doesn't matter) you get a massive decrease of overall scene light using the DMC in comparison to using the Subdivision.
This does not happen when using normal VRay lights.
Any even funnier is the fact, that once you activate GI - results between DMC and Subdivision are suddenly pretty much the same.

So what am I missing?
I don't want to relight scenes when I decide to switch samplers.

Thx in advance,
Flo

[3DMK]

Hi!

This is a thing that has kept me wondering for a while now, and it would be great to finally get an answer on this issue.

Has anyone else noticed, that when you light your scene with a simple dome light (HDRI or no HDRI doesn't matter) you get a massive decrease of overall scene light using the DMC in comparison to using the Subdivision.
This does not happen when using normal VRay lights.
Any even funnier is the fact, that once you activate GI - results between DMC and Subdivision are suddenly pretty much the same.

So what am I missing?
I don't want to relight scenes when I decide to switch samplers.

Thx in advance,
Flo

Are you using sub-pixel mapping???

[FSGFX]

Yes i do.

This is the reason?

EDIT:
Indeed - this solves this issue, so thanks for the tip.
I would really appreciate if anyone had the time to come back and explain, why this is happening with the dome light.

Cheers,
Flo

[vlado]

When you turn on the "Sub-pixel mapping", V-Ray prints a warning that you may get incorrect image brightness - it is true

Imagine that you have 4 samples in the pixel, and for each sample the dome light has intensity 0.25, 1.5, 0.3 and 1.6. With sub-pixel mapping off, these are averaged together for a final result of about 0.9=(0.25+1.5+0.3+1.6)/4. However, if you turn on sub-pixel mapping and clamp output for example, you will get (0.25+1+0.3+1)/4=0.64 which will be markedly darker.

The extent of this darkening depends on how exactly the samples are distributed between the image sampler and the dome light, ranging from negligible to quite obvious.

Best regards,
Vlado