OctaneRender® 1.024 beta 2.47 (lin/mac/win) [OBSOLETE]

[r4ytrace]

man I cant wait for 32 bit version - I love me some caustics :p

It's already there

Cheers,
Marcus

yay cool - gonna check it out tonight

[kubo]

some more tests, dunno why but this one has proven a tough one to solve cleanly.
PMC and PT glass has 0.7 opacity and 0.012 disspersion.
It could be the wood table's bump, I'll try with a flat white later.
Any ideas?
update:
the images are actually in inverse order timewise, I've tried using a white diffuse material for the table (3rd and 4rd shots) and later I've change the disperssion values of the glass back to 0 (1st and 2nd) it seems without dispersion it looks better, and also with a lower iso and higher fstop values it "hides" somewhat the noise.
Added the render kernel values
the last 4 renders have the exact same values in everything except render kernel. Then I adjusted the light emmiters 'cause they were too bright (the next "batch" 1-4) and changed the material for the table, then I changed the glass dispersion, same PMC kernel used in all of them, and PT kept the same values.

[abstrax]

some more tests, dunno why but this one has proven a tough one to solve cleanly.
PMC and PT glass has 0.7 opacity and 0.012 disspersion.
It could be the wood table's bump, I'll try with a flat white later.
Any ideas?

Did you render both images in beta 2.47 with the same maximum depth? This is very important for any comparisons, and the PT result differs greatly from the PMC result. What was the maxrejects you were using PMC?

Cheers,
Marcus

[andrear]

simple interior scene only lighted with sunlight.
1300 samples in 6 hrs on my laptop at approx the same rate of pt.
very good indirect light and nice reflections never achived before with pt.

thank's you guys for the improvement .....

[Refracty]

Cool... Finally:)

The new Release seems to need a muchhigher exposure number to receive the same lighting result. Please 'level this down' so that it is closer to the direct lighting preview engine.

What do you mean? Could you please post an example?

Cheers,
Marcus

I have opened an old 2.46b scene in the new PMC build and had to increase the Exposure Value (+10) to receive a similar bright image (both PT and PMC).
Unfortunately I can't post the image because of the clilent's NDA.
Did anybody make the same experience?

Cheers
Refracty

[romrom74]

Hi there,
I'm testing this test release and experiencing some odd interface behaviors:
render doesn't update when I move any sliders of the pmc/mesh render kernel node, I have to restart the rendering with the button next to the pause button to actually see changes of my tweakings. This is so quiet uneasy to guess what does what when playing with the new exploration_strenght and direct_light_importance sliders of the new kernel! Am i the only one?

Another question: is this pmc supposed to help to clear fireflies faster, those tthat would come from a Sun lit environment? I m noticing almost no difference between PT and PMC with sun light (cast trough glass-less windows, just openings), fireflies-wise I mean.
I may post soon some images of what I m talking about.
Cheers,

EDIT: ok ,concerning the non refreshing render, the thing happens when loading a pre-2.47 tets scene apparently.Loading a mesh in a fresh new scene doesn't show this strange GUI/render problem...

[ycarry]

Test with 3mm diameter Brillant (Diamond standard cut - Refract 2.418, Dispers 0.044) + simple low emitter (dist. 30 cm top back)

PMC 225 samples, 20min:

PT 225 samples, 4min12:

PT 1340 samples, 25 min (:

Pathtracing is really faster than before; and its my prefered for instance as effect in object is near same as PMC
even after 25 min caustic cannot be same; but I will test brillant included in gold jewellery...

Edit: no post process, brut from Octane.

[kubo]

I've been doing some test with a model with a light source and black background. What I don't understand is why with pmc the background gets affected by noise but with pt and dl it doesn't.

[ROUBAL]

Very nice rendered images here ! I can't test for my own currently, because I need my computer all the time.

When I see that Cuda 4.0 is required, I am very worried : I am still stuck with Cuda 3.0, because drivers for Cuda 3.2 didn't allow Octane to see my external graphic cards in my Cubix box. I was wondering about purchasing two Gainward GTX 580 3GB, but I must be sure that I will be able to use them with the Cubix Xpander, because currently I can't purchase one more computer to host them !

[DayVids]

I have seen a decent speed increase on my GTX 260 216 core card. I think this has to be because of a change in the driver or the new cuda, because Frames in 1920x1080 that were taking me 2-3 min. to render in a turntable animation to create a rotating logo, are now taking ~90 sec. I would think that the renders were probably taking between 2 and 2.5 min before so it's a really pretty reasonable improvement. This particular one was being done in DirectLighting, because PT was unnecessary.

Thanks a bunch guys!