V-Ray Image Sampling Question

[dushyantagarwal]

Hi there,

 

I have a question about V-Ray Image Sampling and I would really appreciate it if someone could help me with an explanation.

 

Theoretically, I feel I understand how the image sampling is working within V-ray but the one setting that leaves me a bit confused is the 'Adaptive amount'. From reading on the forums, I understand that if we set the Adaptive amount to 1, V-Ray will only use the adaptive minimum samples initially which might not be enough for V-Ray to estimate the noise correctly in some cases like occlusion and off course if I set the Adaptive amount to 0, it won't work adaptively at all. But I'm a bit unclear about the approach to judge when an Adaptive amount of 1 would work well and when turning it down will improve results.

 

I read that one would really benefit by turning down the Adaptive setting lower than 1 in case of the VrayDirt Shader for ambient occlusion so as to enable it to take more samples from the subdivs specified within the dirt shader.

 

In a recent video tutorial I watched, it was recommended not to keep the adaptive amount to 1 in case of using HDR's but I don't quite understand why.

 

I apologize if the scope of my question is too vast or if I was a little vague but any clarity on this subject would be really helpful.

 

Thank you.

 

Dushyant

[Brian Smith]

Adaptive amount controls the % of adaptiveness for settings where adaptiveness is possible. At 0.0, the noise threshold is ignored and all adaptive settings become fully brute force. At 1.0 all adaptive settings become 100% adaptive. The higher the value, the more the algorithms become truncated...the lower the value the longer they are allowed to compute a better solution. You should always set it to 1.0 unless you know that the noise you see somewhere can only be reduced by reducing this value...which tends to be the case with certain areas of vray like HDRIs, and possibly VRayDirt (which I haven't personally noticed). Reduce to something like .9 or .85 in these situations. All other times, leave it at 1.0 to give vray the opportunity to be more efficient than brute force.

[dushyantagarwal]

Hi Brain,

 

Perfect, Thanks for explaining that further. So in a situation where I still see noise in a certain reflective or refractive area for example, would lowering the adaptive amount and increasing the subdivs for the reflection/refraction parameter in that particular shader be the correct approach instead of say lowering the overall threshold value to begin with?