OctaneRender® 1.025 beta 2.51 [OBSOLETE]

[Mugga]

I found a bug with the HEX-Code for RGB-Spectrum.

The HEX-Code in 2.51 is different than in 2.49b. I copied a code from 2.49b to 2.51 and the color is completely different.

[abstrax]

I found a bug with the HEX-Code for RGB-Spectrum.

The HEX-Code in 2.51 is different than in 2.49b. I copied a code from 2.49b to 2.51 and the color is completely different.

The colour as it is displayed in 2.51 represents the colour as it will look like when rendered, due to the gamma correction, which was not taken into account in the old colour picker.

Cheers,
Marcus

[abstrax]

Thanks for the betas Octane team. I'd like to share some issues I'm running into.
1. In Linux, when I launch octane, all of the settings from the right panel are moved (mismatched) 1 place; i.e. in the screenshot I'm mousing over camera options and it displays "current resolution" options, If I click on current resolution options it will display the environment options and so on:

Only after maximizing/minimizing the interface do they go back to their respective positions.

That sounds like a bug in the UI toolkit and it should happen in older releases, too, since we haven't changed anything in that area.

2. In Linux, I load an obj file, but scene's tris/gpu count displays 0 so all I get is an empty/gray scene.
3. In windows, the obj file loads fine, but after tweaking the scene options a bit the render freezes. All the rest of the interface seems to be alive, but I can no longer orbit, rotate, pan or zoom my model in the viewer.

Gonna run some more tests to see what else I run into.
cheers,
Alvaro

Which kernel did you use? Could you send us the OBJ file, including textures?

Cheers,
Marcus

[abstrax]

I notice that the tool tip problem with the 2x2 and 4x4 sub sampling icons have not been solved yet.

That's an old bug that exists since I don't know when. We haven't spent much time on investigating it since it seems to be a problem deep in the UI toolkit.

Cheers,
Marcus

[abstrax]

with 2.48 - 2.5x i had the same problems as felix. especially with huge scenes (about 2gb vram usage).
if i remember right navigation was much better under 2.46. atm i have no time/hardware for specific testings., maybe next days i find some more details and could post some test scene.
@marcus: could you use a blend file? (obj/ocs will bee to big)

hardware/os: see sig amdX6 sys +mint10

Yes, a Blender scene should be fine as long as it includes all textures.

Cheers,
Marcus

[abstrax]

I have some heavy performance problems on Linux (Kubuntu 11.04, driver version 280.13), the view updates are VERY slow, sometimes I can't even pause a render and it's impossible to do anything else while rendering, because there is a view update just every second or so. Later in the rendering process (after several dozen or hundreds of samples) it's even worse, I just had 5 minutes to kill octane, as the system didn't respond really on any clicks anymore. The whole screen was more or less frozen, the only thing possible was moving the cursor.

The problem already appeared in 2.49b and 2.50 (don't remember all builds before, but it worked once perfectly for sure)

We tried to solve/reduce this problem in beta 2.50. Didn't it get any better from beta 2.49b to 2.50? How the viewport navigation? Is it better than the UI behaviour?

We think we know where the problem is, but don't know exactly how to solve it, since it is an issue in the X server, which seem to behave differently on different systems...

Cheers,
Marcus

[acc24ex]

this one runs OK on my setup of 2x580 590 480gtx on a simple sphere test.. The system UI is 100% responsive on PMC kernel - as it was on the previous few versions, so just making a note of that in case others didn't experience it
..

Anyway can we get a bit more explanation on how is dispersion coefficient b, absorption color, depth of dispersion supposed to behave on specular so we know how to use it and for what exactly (examples might come in handy)

And also on the pmc kernel, how is direct light importance, exploration strength, maxrejects supposed to behave, and some basic explanation in which cases what numbers to change..

[kubo]

I'm doing some tests right now, and it runs great, before vacation I had a constant crash on 2.49, I didn't have time to check it out, but yesterday I did some tests. I had been swapping GPUs and it seems that my secondary 570 was too close to the mobo's speaker pins and the cable had stuck one of it's fans, so it's temperature was going up too fast and crashing the GPU, go figure, I've swapped back the GPU's and Octane is running smooth as ever. But now with more tricks, hehehe, time to do some fun testing.

[Draydin_r]

Hi Guys,

I've noticed a caustics problem with 2.50 and 2.51. I have provided the examples below.

The 'test head' scenes are identical between 2.49b and 2.50. Any ideas of whats going on?

[roeland]

Yes, 2.50 and 2.51 contain a bug in PMC, preventing it from rendering caustics properly. This will be fixed soon.

--
Roeland