Controlling texture quality?
-
Is all this without Filter Blur?
-
Yes, without Filter Blur.
Now, I switched back to Adaptive DMC, min. 1 max. 16, but with noise threshold and color threshold to 0,001. Kind of extreme values for thresholds, but it's almost flawless, and the rendering time in good shape, 8.30 minutes.
I think if you have some texture, the noise will blend nicely and you'll be fine with thresholds @0,005. Here is the image
-
These tests are actually quite interesting.
Too bad I'm home now. Don't have my V-Ray set up.
-
-
i agree with stefan, using adaptive dmc with lower treshold will solve the matter. and also ticking ambient occlusion will further bring out details..
-
I always use a little bit of AO.
I found that a little Filter blur to the TexBitmap helped greatly and reduced the need to increased adaptive DMC too much.
-
@thomthom said:
I always use a little bit of AO.
I found that a little Filter blur to the TexBitmap helped greatly and reduced the need to increased adaptive DMC too much.
honestly its my first time to see the effect of filter blur, i am interested to find out more by doing some further testings..
-
I don't think 0.0 has used to be the default. In Max, with Mental ray it's 1.0 (which I think is too much though.)
VfSU's defaults isn't the best. Max' defaults are better tuned IMO. Anyone with V-Ray for Max - can you check what the defaults are there?
With Filter Blur set to 0.0 it'd be no anti-aliasing - hence the noise.
-
I never touch the filter blur. All of those broken lines are controlled by your Image Sampler. Some say that the image sampler is the most important setting for quality.
This video explains the fixed rate sampler and how your settings clean up an image but translate to longer render times.
-
Nice video. Might have to check out the rest of that.
So, how do you work with the sampler? Do you leave minimum at 1 and tweak the Color Threshold until satisfactory?
-
But in terms of the model I posted, the texture has black and white lines - the most extreme color values you can get. You'd think that was enough contrast for the threshold to trigger higher samples?
Advertisement