Caustics and fog dont render on reflection and refraction.

[ior]

Is this infinitely small intersection points and infinitely thin ray paths can be stored on disk. updating at a click of a button? No mater how time it takes. No more than we take on render!

[lycium]

the method we're working on should do a great job

[pixie]

the method we're working on should do a great job

And once GCC gets used we the windows users will get a major speed boost!

Updated: bust->boost, many thanks to Borgleader for having noticed it, we want speed to go free!

[Borgleader]

the method we're working on should do a great job

And once GCC gets used we the windows users will get a major speed bust!

bust --> boost

But yes, major speed increase is going to be very welcome!

[StompinTom]

Hungry for speed here! Can't wait to play with this new toy!

[ior]

I see that developers moved a step towards the possibility off rendering refraction caustic effects by using Glass Accel.
Right now, in version 2.6.1, there is still a problem with multiple refraction objects (glass windows or whatever glass object), the second object that refract the caustics is still a little bit darker than the first and subsequently the rest of the objects are even darker.

Note that you can only use Glass Accel with sun and no exit portals.

Will this be available for other emitters as well?
I think the aim is to create a reflection acceleration method also. When can we see this implemented?

[Kristof]

I see that developers moved a step towards the possibility off rendering refraction caustic effects by using Glass Accel.
Right now, in version 2.6.1, there is still a problem with multiple refraction objects (glass windows or whatever glass object), the second object that refract the caustics is still a little bit darker than the first and subsequently the rest of the objects are even darker.

Note that you can only use Glass Accel with sun and no exit portals.

Will this be available for other emitters as well?
I think the aim is to create a reflection acceleration method also. When can we see this implemented?

I have recently finished a test using only mesh emitters where glass acceleration dramatically improved the results, you can refer to this link to see the test.

[Kristof]

OK here is my reflected caustics test, 30 minutes with a laptop.

I am also making use of glass acceleration. I have yet to run any comparisons with and without as of yet.

[ior]

With glass acceleration Indigo really rocks!

Can't wait to see it applied to other emission objects.


Here are some examples:

with glass accel

without glass accel

with glass accel and mirror

In the last image see the perfect caustics of the pot in the mirror!


I modeled the room opening badly, that's why on the room opening, the bottom is illuminated by the sun, on the side there are reflections of the right side glasses.

[Kristof]

With glass acceleration Indigo really rocks!

Can't wait to see it applied to other emission objects.

I modeled the room opening badly, that's why on the room opening, the bottom is illuminated by the sun, on the side there are reflections of the right side glasses.

Once more, glass acceleration works with mesh emitters and other emission objects as well as with the sun.

Also here is the same test without glass acceleration, this time only 14min though.

[ior]

Once more, glass acceleration works with mesh emitters and other emission objects as well as with the sun.

It does, didn't realize that because the tests I've done, were with small light sources.

Glass acceleration of mesh emitters seems to work better on refraction of caustics, small light sources still produce fireflies difficult to converge.

Why the floor is darker in the mirror side of your image? It is because the caustics are not being reflected or your mirror isn't near 100% reflective?


When you use the sun you cannot use outside exit portals for glass accel to work.
Volumetric fog is still not reflected or refracted with glass acceleration.

[ior]

Here is an example of fireflies difficult to converge with glass acceleration and a small emitter.

small emitter fireflies

there is a sheet of glass in front of the camera with the edge ending in the middle of the pot caustics.

The fireflies will never converge in an acceptable time (24h).

[Kristof]

Hi ior,

Are you using MLT? It is near impossible to work with caustics without it.

[ior]

Are you using MLT? It is near impossible to work with caustics without it.

Hi.
I didn't try MTL, thanks for reminding me, maybe MTL also got improved with glass accel, I'll make some tests.

[Zom-B]

...maybe MTL also got improved with glass accel...

MLT owns every scene that uses reflections, refraction & Co!!!
Also Try MLT vs MLT+BiDir!
With BiDir the Samples/s should be lower, but their "quality" is better, so you maybe end up with a better result.
Depends from scene to scene...