[Kabe]

...I don't know if there's any possible or not as I don't know much about Maxwell's render method.

Anyway, I think it's pretty obvious that certain areas of an image need far more samples than others.

The affected areas are already visible after a few minutes. Is there any chance to increase sampling for these areas? Either using a mask to control render density or an automatic mask that looks for local "noisiness"?

Keep up the great work

Klaus

[Thomas An.]

Yup, and I was thinking about this the other day... maybe something like making a window selection (live) on the display window (the same way we select regions with NeatImage) and have Maxwell concentrate all CPU power on that spot only (with some feathering around the edges) ...

[John Keates]

Might it be possible for MR to analyse the render to see how much noise there is so that it can concentrate its resources automatically? This should be possible and would provide particular advantage when rendering animations.

[Kabe]

It may be easier to shoot more samples when MR encounters a material that leads to noise like a dieelectric with low Abbe, or a rough metal...

Just a thought

Kabe