New progressive caustics problem

[Jiri.Matys]

It's just a bit odd....I just tried the same thing with the simple scene and without caustics, and with the noise at 0.01 it took 1.11 and with the noise at 5.0 it took 52 secs.
That's just not how I imagined it worked and is such a large difference that, yes, it's a good thing to have noticed and could speed stuff up (if it scales), but is pretty unintuitive
and extremely unexpected.

EDIT: I put this together just now. With noise at 0.1 I stopped it as it was not going to finish - time 5.x
Upped the threshold to 5.0 for the second one and it finished as you can see in just under where I randomly stopped it for the first.
The noise is obviously a little different but overall the quality is the same.

Hey, fixeighted, how you light the scene? sun sky?

[fixeighted]

Yeah sun and sky...

[Jiri.Matys]

Yeah sun and sky...

What about hdr?

[^Lele^]

It's an issue with the way the tech works right now, not with how the scene is lit.
For now, the best one can do is follow the doc's tips.

[fixeighted]

hdr is a mess in this instance....just to answer the question.

I did follow the suggestions to the letter with the previous test though.
I'm sure it will improve.

[^Lele^]

Could i get the scene?

Anyhow, both aspects will be looked into.
For now, the best you can do if a scene misbehaves so, is render it as fixed sampling (i.e. same min and max aa.).
Noise threshold is then never checked for, so long as min AA hasn't been reached.
If you feel adventurous, fiddle with the ray bundle size, although the margins there are small to non-existent.

[^Lele^]

I should also clarify a couple of things about caustic generation.
IBL HDRIs will contain often sizeable lights, and those will produce softer caustics, which will in turn require (many) more photons to clear up.
The smaller the light source (from spot and point ones to area, disc and sphere and IES ones), the sharper and quicker to resolve the caustics will be, conversely as the sizes grow.
Analogously, the rougher the specular or refraction, the more diffuse the caustic, to the point the engine will stop calculating it and just treat it as a diffuse bounce.
The first, an IBL with huge windows, had to have caustics multiplied by 10, and still was looking like that after 8 minutes.
The second, a small spherical light, was at that point in two and a half.

[Jiri.Matys]

I delete my test scene by mistake. I want to go back to my original post in this thread and want to try this one more time, where is a problem. I follow docs tips so really dont know what is wrong. But I lit a scene with VRay light if I remember correctly.

[^Lele^]

We'd be happy to check anything you have to share out.
While we surely have ideas of what needs improving (well, Dian and Vlado do.), we can also do with scenes which misbehave.
It is not a bit-identical ripoff of Corona's, it has parts insipred by one of their papers, but the rest of it is either novel or very different, and is implanted on a very different engine architecture, so we would be grateful for stuff which could be profiled for bottlenecks.

[Jiri.Matys]

So I dont make the test scene but now I want to use them on regular interior scene. Caustics looks ok, but I have a problem with terrible noise in a mirror. This is a caustics pass after one hour Where is a problem, what can I do? I think VRay wont denoise that.

[Jiri.Matys]

Here is the picture, sorry.

[super gnu]

We minnows, with only 8 cores, never quite get to see the problems of you biggies with 64, eh.
But i can imagine how being stinking rich (in cores) might be.

ohh its wonderful, although sometimes i have to be a bit tough on the kids i have in the hamster wheel, they have a habit of slowing down every so often, and with 64 cores, they need to keep the pace up or the lights go out, my *extremely* raytraced render fails, and the champagne in the fridge gets warm. pff.

[Paul Oblomov]

I have 7% CPU utilization Both progressive and classic. I remember I mentioned this at the closed group (btw lele - what's happened to the group? half a year and can't find a new place maybe you know where it moved).

[^Lele^]

LMAO, that's a record allright, Paul!
I have actually not entered the group in ages (i'd use the mail digest to read and reply, often *not* to all.), so i'm not sure.
This ought to be the new place for it, but am not sure what could be done for the old content.