PhotonGI cache

[lacilaci]

The brightness of the benchmark scene with PGI is like PT without PGI

not if you use a lot of paths, like 100. Then PT is brighter!

What do you mean with "paths" ? The max. path depth ?

Yes, path depth for all rays (total glossy diffuse... all of them)

[provisory]

The brightness of the benchmark scene with PGI is like PT without PGI

not if you use a lot of paths, like 100. Then PT is brighter!

But bidir is still much brighter. I've tested it earlier (before PGI), and PT was always noticable darker than bidir.

[lacilaci]

The brightness of the benchmark scene with PGI is like PT without PGI

not if you use a lot of paths, like 100. Then PT is brighter!

But bidir is still much brighter. I've tested it earlier (before PGI), and PT was always noticable darker than bidir.

Yes... bidir is brightest, then PT and high pathdepths and lowest brightnes you'd see with PT+PGI...
Maybe it is expected behavior, but what's interesting to me is, that in a very basic scene PT+PGI will be even brighter than bidir... so that's why I think there could be something wrong. One way or another, I don't mind a little bias for the huge boost in speed, but still feels off in some cases.

[lacilaci]

I always keep thinking that textures (just as hdri) somehow kill cache... I don't see other reason for weaker lighting from PGI if with simple materials it just works... maybe something with gamma settings or so...

[B.Y.O.B.]

PGI look So dark on Nakagin capsule scene . Or don't have a given lastest fix

I have created a simplified version of this scene.
However you didn't post the render config settings you used, so mine may be off.

The scene consists of a cube with a hole, lit by an HDRI as only light source.
Included are the .blend file and LuxCore cfg/scn exported via Filesaver.

Note: I forgot to disable autolinear tonemapping, see this post for a fixed scene: viewtopic.php?f=5&t=840&p=10565#p10565

[Sharlybg]

I have created a simplified version of this scene.
However you didn't post the render config settings you used, so mine may be off.

Thank you for that effective solution.

[Sharlybg]

Also hope the fix for this will be in 2.2 alpha 0 as it is really obvious difference.

[Fox]

Does the total number of photons have to be the sum of indirect and caustic?
What units are the glossy T, the values go from 0 to infinity?

[Dade]

Does the total number of photons have to be the sum of indirect and caustic?

Nope, it is the number of photon paths traced, each path can create 0 to many cache entries.

What units are the glossy T, the values go from 0 to infinity?

The are the U/V roughness parameter of glossy and glossy-like materials (i.e. [0, 1] range).

[frank_yifei]

maybe the brightness difference comes from tonemapping in camera setting, just guessing.