OctaneRender™ Standalone 3.06 Tweet

[carterner]

Can octane light add adaptive sampling features.because if there have many light in a scenes ,how much the sampling for each light ofen difficult to be sure .can octane make the light sampleing automatically in the future

[andw]

Looks like there was some breaking changes in 3.05 -> 3.0.6 work with opacity maps: I tried the 'Materials/Non-organic/Cloth/Roman shades' material from LiveDB, in 3.05 it renders well, in 3.06 the surface looks completely transparent (even if I render only material on test ball). So, is it a bug in OctaneRender or just that material needs some fixes?

[abstrax]

please, pretty please, add autosave exr feature

Autosave the Octane file and autosave your render at x intervals. Really needed feature.

HI, otoy dev team, would u guys please don't ignore our request please? autosave is really important feature that several engines are already have, but not in octane, or should I say you guys are never never never losing rendering process during any kind of issues like gpu too heat or crash? even the new vray 3.5, they already support autosave and resume feature, same to corona render. so you guys just hit the render button and let many gpu running, finally you always get what you want and never had crash or any issue that cause losing render process???

We are not ignoring it, it's just that we didn't get to it yet. There are many feature requests we haven't implemented yet. It's a bit more complicated than you think, because saving the render film isn't enough. To reconstruct the full render state, you also have to store the project that is currently being rendered - but only if it was changed. But if you use the batch render script, the original project isn't even touched, because the rendering is done on a copy of the project. And if it dies on a specific frame, how can then rendering continue at the time it died, etc..

In a nutshell: We really don't know yet how to implement it so that all cases are covered. To figure that out, takes a bit more time which we didn't have yet.

[abstrax]

apologies if i repost this same question here from the development forum, i found out only now that 3.0.6 is final, but have to wait until it is supported by the LW plugin.

i did see reported this already somewhere else, just not sure if that has been fixed: when using the PMC kernel, on my networked windows machine with 5 x geforce 980ti GPU's connected, i can see that the GPU activity never goes beyond 75% on the single GPU. in fact, the fans don't even make any noise when rendering that way. totally expected behaviour when rendering with path tracing: 100% GPU usage and quite some fan noise! has this been fixed with the latest stable update? using octane with the latest lightwave plugin, here...

cheers

markus

Yes, I answered that one multiple times already: The problem is that modern GPUs produce samples faster than they can be splatted in the film buffer. We are working on tone-mapping that runs on the CPU, which will then allow us to structure the film buffer differently which should improve splatting speed. At the moment the tone-mapping is done on the GPU, which requires a specific arrangement of the film channels, which doesn't work very well for PMC.

[abstrax]

Looks like there was some breaking changes in 3.05 -> 3.0.6 work with opacity maps: I tried the 'Materials/Non-organic/Cloth/Roman shades' material from LiveDB, in 3.05 it renders well, in 3.06 the surface looks completely transparent (even if I render only material on test ball). So, is it a bug in OctaneRender or just that material needs some fixes?

It's a change in the behaviour of the image texture nodes connected with a greyscale texture pin (like opacity): Until 3.05 image textures were applying power first and then invert when connected to texture pin that expects a greyscale (like bump, opacity, roughness, etc.), but applied invert first and then power, if it was connected with a texture pin that expects a colour (like diffuse, specular, etc.). This is not only inconsistent, but also had weird effects, for example the texture preview was showing a different result than what actually was used.

When we merged the various texture fetch functions into one, we noticed this issue and fixed by picking one way and then using it for all texture fetch modes: Invert is applied first and then power. This breaks the mentioned material and to recreate the old behaviour, you have to disable the invert option in the "SEFA mask.jpg" node that is connected with the opacity pins and feed this texture into an invert texture node first and the invert texture node into the opacity pins.

[3dworks]

apologies if i repost this same question here from the development forum, i found out only now that 3.0.6 is final, but have to wait until it is supported by the LW plugin.

i did see reported this already somewhere else, just not sure if that has been fixed: when using the PMC kernel, on my networked windows machine with 5 x geforce 980ti GPU's connected, i can see that the GPU activity never goes beyond 75% on the single GPU. in fact, the fans don't even make any noise when rendering that way. totally expected behaviour when rendering with path tracing: 100% GPU usage and quite some fan noise! has this been fixed with the latest stable update? using octane with the latest lightwave plugin, here...

cheers

markus

Yes, I answered that one multiple times already: The problem is that modern GPUs produce samples faster than they can be splatted in the film buffer. We are working on tone-mapping that runs on the CPU, which will then allow us to structure the film buffer differently which should improve splatting speed. At the moment the tone-mapping is done on the GPU, which requires a specific arrangement of the film channels, which doesn't work very well for PMC.

thank you for the answer!! hopefully that can be fixed, PMC is great!

kind regards

markus

[PolderAnimation]

We have a scene with around 150 'object layer' nodes. This gets extremely slow in the scene graph. The computer has an hard time displaying them.

Then an other question would it be possible that the 'Material Map' node also works on alembic caches (and groups)?

Cheers,
Sander

[Jorgensen]

hi abstrax

i have attached the log.

jorgensen

[Phantom107]

When I hook up a Scattering medium to a Diffuse material it does not do anything in Path Tracing, even when I play with the settings. Seems to work fine when I hook up the same Scattering medium to a Specular material.

Am I missing something?

[calus]

When I hook up a Scattering medium to a Diffuse material it does not do anything in Path Tracing, even when I play with the settings. Seems to work fine when I hook up the same Scattering medium to a Specular material.
Am I missing something?

Are you sure you didn't forgot to increase transmission of the Diffuse material ?