OR since v3.00 is caustics later than v2.25 at same S/px

[nuno1980]

Are you still investigating this issue? New v3.0x or v3.1x?

[nuno1980]

This issue list isn't for v3.03.4!?????????????????

[nuno1980]

No any response...

[nuno1980]

I'm worried because the response is too slow. Therefore, the price of OR may be BAD.

[nuno1980]

@roeland: Are you working in progress to fix this issue?

[nuno1980]

Almost 2 years later... Still no fix... I tested OR V4RC1 but nothing fix again yet...

OR v2.25 is normal - PMC@44 diff-,@82 spec-depth,@10k rejects; 120k spp - 32h17min running my EVGA GF GTX 980 Ti SC+.

--- OR v4 RC1, v3.0x and early v4:

Few caustics are BADLY missed without "caustic blur".

Seems to be normal at caustic blur 0.01


-----> Can you find any problem of PMC kernel of OR v3/early v4? 10k-reject is wrong as bug because I think only 1000-reject (this number HIDDEN of rejects), ok?


--- Download ORBX --

carefes.orbx

(56.95 MiB) Downloaded 411 times

[nuno1980]

If we have the licensed customer -> URGENT: PLEASE must reply!!!!!!!! Do NOT delay more than 2 or 3 days, if you use PMC kernel with multi-top-end GPU cards!!!

[nuno1980]

New image rendered

It's normal in v2.26.1 -> See difference of images rendered between this new image and «Few caustics are BADLY missed without "caustic blur"» (Thu Aug 23, 2018 9:01 pm UTC).

Same caustics = same render time but v3/early v4 is more ~2.5x number of samples per pixel than v2 but v3/early v4 loses/misses few caustics.

[FAZ]

Yo homie!

Why not use Pathtracing?

6000 samples
diffuse = 4
Caustic Blur = .01
GI Clamp = 700

Make sure "fake shadows" is OFF.

If you want to get rid of more spots, lower "Hot Pixel Removal". Using "dispersion" in clear materials makes this almost impossible.

The way your liquids are sitting inside of the glass looks wrong. Check make sure dielectrics your model is correct for this kind of thing. See...
viewtopic.php?f=101&t=68220

[nuno1980]

3.5 years later...

New OR 2022.1 XB1 with new Photon Tracing kernel and AI Denoiser!

OR 2022.1 xb1 - new photon tracing.png (122.87 KiB) Viewed 3128 times

Thank you!