I seem to recall at one point someone mentioning that when you had a lot of glossy reflections/refractions, motion blur, DOF, etc, that adaptive image sampling wasn’t necessarily the fastest way to go.
Is this true? Should I be using the fixed aliasing for such effects to speed up rendertimes? If so, what would an equivalent fixed sampling amount be to an adaptive of min 0, max 2?
Any help would be appreciated as I’m trying to speed up rendertimes with 1.4570.
I wish I could find the post but it was Vlado who said it was faster when there are a lot of blurry effects.
Setting of 1,4 is pretty close to what you are used to with adaptive it’s faster but a little jaggier.
Today Vlado described some of the changes since the old style:
The Multipass option is always ON in 1.46 since it made little sense to turn it off.
The Rand option is always ON if the number of subdivs is more than 1.
The Threshold value is taken from the QMC Sampler’s Noise threshold. If you do not reduce the default value, you will always get to the point where more subdivs in the material/light etc do not reduce noise anymore.
The reason why the AA threshold and the QMC Sampler’s Noise thershold are linked is because they are part of the same QMC sampling system; if they are different, it is very easy to mix them up in an inefficient way.
Best regards,
Vlado
This was interesting because settings higher than 1,4 seemed to make no difference, which seemed odd but I guess a lower noise threshold should make a difference. Need to test it.
hmm, I thought I remembered it to be simple two level but it was fixed .
If you have lots of glossy effects, it’s best to use Fixed rate. 3 subdivs is probably too low for the fine detail that you have; use 4 or 5. Also try playing with the QMC sampler parameters to control the quality of glossy reflections (for example, increase the Min samples, or lower the noise threshold and/or the early termination amount).
The equivalent of 0/2 rate for the Adaptive subdivision sampler is 1/4 for Adaptive QMC sampler, or simply 4 subdivs for the Fixed sampler.
now THATS a tasty morsel of information. Have always wondered that. Is there a mathematical formula for figuring that out? We use adaptive qmc 0,3 for animation here at work but are still getting some noise over the animation. Others thought it was AA, I thought it was qmc settings. We’re trying to get that resolved before we start switching from NTSC to HD for some projects.
The noise may be many things - a (bump) texture with small details, glossy reflections, area lights/shadows etc. If it is a texture then you need to increase the AA settings (increase the Max subdivs and/or lower the QMC Noise threshold). If it is a glossy reflection/area light/area shadow you can:
(a) increase the corresponding subdivs; reduce the QMC Noise threshold; increase the QMC Min samples.
or (b) increase the Global subdivs multiplier; reduce the QMC Noise threshold; increase the QMC Min samples.
I myself prefer (b) since I don’t have to go through all the materials/lights/shadows etc and adjust individual settings. Of course, individually adjusting the subdivs allows greater control over rendering times.
thanks for the info. we’ve seen the noise on exterior animations where there are no glossy reflections or vray lights, but we somehow got into the habit of increasing the QMC noise threshold alot. We’re looking for a good time to switch over the the 1.45.xx so theyre will be alot to adjust to.
