Multi-region better for more right caustics

[00Ghz]

There is definitely a difference between the 2 images. The multi region one has brighter caustics in the interior glass.

[nuno1980]

There is definitely a difference between the 2 images. The multi region one has brighter caustics in the interior glass.

Yes but caustics in ground - look upper center of ring.

multi-region is better for highest priority for caustics. lower resolution of region - not = higher priority but better ballance for resolution of region ~1002x602.


search better caustics at same number of samples per pixel...:
-- lower resolution of region - higher priority of caustics at high number of rejects because caustics are less noises, great and closer but right caustics are later at low number of rejects.
-- better ballance for resolution of region ~840x480 or 1002x602 between high number of rejects and low number of rejects.
-- higher resolution of region - lower priority of caustics at high number of rejects because caustics are more noises, little and far but right caustics are earlier at low number of rejects.

Can you understand well?

[nuno1980]

I edited to add "rejects" in previous message, sorry.

[nuno1980]

New release of OR standalone??

[abstrax]

Currently it's not planned to implement multiple render regions. For path tracing and direct lighting we plan to implement adaptive sampling which should help with your type of scenes. For PMC it's quite tricky to make it work for non-rectangular regions and multiple rectangular regions. Roeland has an idea he would like to try out later, but he is currently working on OSL so that would have to wait.

[nuno1980]

Currently it's not planned to implement multiple render regions. For path tracing and direct lighting we plan to implement adaptive sampling which should help with your type of scenes. For PMC it's quite tricky to make it work for non-rectangular regions and multiple rectangular regions. Roeland has an idea he would like to try out later, but he is currently working on OSL so that would have to wait.

Thank you. If you're working in progress for multi-region at PMC during long time then I want unlimited samples per pixel in next release.

[Goldisart]

tell me how IRAY caustic forms//// it would be great to animate the fake caustics on the principle of action IRAY /// I work a lot with the jewellery and if you have the support of diffraction it is necessary to do something with caustical for the animation

...........
as long as the option is just to bake the caustics map and rotate the camera around the object ... but caustic is dead still




but baking caustics with physical it is also quite a long process

[Goldisart]

already two years have passed and the problem is not solved with a caustic - the other day I went to the redshift site and was surprised ... I understand that it uses a completely different type of adaptive .... but render speed 25 seconds, I was surprised at the GTX 660

[nuno1980]

@Goldisart: Thank you but Redshift has only up to 10 depthes while OR standalone up to 2048 depthes at PMC kernel.
Iray cannot make or makes very hardly indirect caustics at glass without roughness or bit roughness.

I do NEVER like any fake caustics!!

Therefore, I prefer OR!

[Goldisart]

Yes, PMС renders the correct caustic ... but if you animate a video caustical it is virtually impossible

............. for animation you need to use a fake caustic