OctaneRender™ 2019.1 RC1 Tweet

[Zay]

How were you able to test? If another machine you may have introduced whole number of other variables.
Anyway, I was there myself not too long ago, I cloned my win 7 drive and updated a copy. Never went back to the 7, too many other apps now using its feature set.

Did the usual. Move cards around, changed bios settings, new drivers, installed a clean win 7. Still no up-sampling. Upgraded to the free win 10 Pro and then I got Up-sampling. But the less Vram and impossible to work in my 3D app while rendering, which I have no trouble with in Win7, I think it will be a long long time before I move to Win 10 and any versions higher than 2018.x.x unfortunately.

[funk]

[*] Fixed toon material rendering black.

This is still broken

Left = Octane 4
Right = 2019.1 RC1

[Notiusweb]

I am finding this version better shows off great possibilities using denoiser + upsampling.
With a 1.5 up-sampling mode also we could enjoy the benefit of up-sampling + denoising + detail, all at fast speeds!
I know dev team said in XB2 thread that 1.5 will put in. But I just want to comment that 1.5 up-sampling will be very good, especially with denoiser, which as expected tends to blur the scene a touch.


PS - I already reported that pickers are broken when using up-sampling. But let me tell you, it is a serious pain in the oz to navigate and test RC1 with a complex archviz scene with up-sampling on, and then have to turn off up-sampling to focus pick, and then turn back on up-sampling, and then want to material pick, and then have to turn up-sampling off, and then change material, and then turn up-sampling on again, over and over again.
If you could do a Hot Fix just for the pickers (focus, material, object, etc...) that would be appreciated indeed.

[haze]

1) Layered material: viewtopic.php?p=364046#p364046
Both materials render the same now, but have become quite dark. I don't think this is right

I think the original was incorrect. Lower your IOR and you should see what you're looking for.

2) Composite layer renders incorrectly: viewtopic.php?p=364054#p364054
Both materials render more similar now, but there are still odd artifacts happening from the composite material.

3) Fast rounded edges are broken using the geometry "merge unwelded vertices" option: viewtopic.php?p=364413#p364413
Still a problem in RC1

5) Artifacts with these universal material settings: viewtopic.php?p=364496#p364496
Still a problem in RC1

6) Octane is not updating when changing rounded edge radius: viewtopic.php?p=364506#p364506
Still a problem in RC1

7) CUDA crash toggling shadow catcher pin on and off: viewtopic.php?p=364534#p364534
Still a problem in RC1

Have made internal tickets for these.

4) Rounded edge has artifacts where corners meet: viewtopic.php?p=364416#p364416
Still a problem in RC1
8) Rounded edge artifacts: viewtopic.php?p=364878#p364878
While the really obvious artifacts were fixed on the small mesh, the big mesh has worse rounded edges than XB2 now
9) Rounded edges: viewtopic.php?p=364887#p364887
Once again, the obvious artifacts were fixed, but I now see some "lines" on what should be a smooth surface. Toggle rounded edges off fixes it.

For the moment, this is a limitation of the implementation. We have plans to improve it.

10) Ai light is making things darker: viewtopic.php?p=364921#p364921
Still a problem in RC1

Let me get back to you.
Edit: seems the fix didn't go into the stable release branch. Sorry about that. It'll be in RC2, which I've just confirmed it is fixed.

[funk]

Thanks haze but I disagree with this.

1) Layered material: viewtopic.php?p=364046#p364046
Both materials render the same now, but have become quite dark. I don't think this is right

I think the original was incorrect. Lower your IOR and you should see what you're looking for.

1.5 IOR is correct for a clearcoat already. I have been running more tests and this appears to be a problem with the dielectric/coating layer, only when the base material has a medium/transparency.

Have a look at my attached example

LEFT = Using a dielectric layer with IOR 1.5. The base material becomes too dark
RIGHT = Custom clearcoat using a metallic layer (fully reflective), then emulating a 1.5 IOR with a falloff in opacity (0 degree facing reflection = 4%, 90 degree = 100%). The base material does not become darker

[divasoft]

Hellow. Attention! Impossible to render with motion blur!
I wrote 2 times about this problem Aktar. He ignores me.
I'm use Cinema4d. When motion blur is turned on in the scene and there are many particles (hundreds of thousands) the time to calculate motion blur increases frame by frame and can reach huge values ​​in a few hours per frame. BUT! If i restart the render and start from the same frame, the render time is reduced to a few minutes. Because of this, it is absolutely impossible to render projects with a huge number of particles. This error is repeated on any scenes, on any computers. I can’t check this bug in standalone, because it is impossible to export particles from Cinema4d to Stand., but I believe that this problem is related to the core.
Briefly: the same frames with a huge number of particles and turned on MB can be rendered for 5 minutes and for 2 hours.

Please solve the problem !!!

[divasoft]

And how i can see, MB calculation uses only one core and total CPU loading 5-8%. Can you make it process multithreaded to accelerate it?

[Jolbertoquini]

Hellow. Attention! Impossible to render with motion blur!
I wrote 2 times about this problem Aktar. He ignores me.
I'm use Cinema4d. When motion blur is turned on in the scene and there are many particles (hundreds of thousands) the time to calculate motion blur increases frame by frame and can reach huge values ​​in a few hours per frame. BUT! If i restart the render and start from the same frame, the render time is reduced to a few minutes. Because of this, it is absolutely impossible to render projects with a huge number of particles. This error is repeated on any scenes, on any computers. I can’t check this bug in standalone, because it is impossible to export particles from Cinema4d to Stand., but I believe that this problem is related to the core.
Briefly: the same frames with a huge number of particles and turned on MB can be rendered for 5 minutes and for 2 hours.

Please solve the problem !!!

We had similar issues with Maya actually is complicate with the plugs I think.

One thing we did to fixe was on the soft side is export our particles as Alembic caches and add at standalone at scene exported or in other shots we did as "orbx proxy" on the scene, this technics allowed us to render really fast a big amount of particles, but also you are completely right we had a project recently with a huge amount of particles from Houdini and was really slow compare to other engines here when we have opacity and higher motion blur settings.

100k particles is not that much for Octane, we did a project here 15millions particles with opacity and motion blur the render was fine like 20min full scene only particles 256 sampels 6min in one 1080ti (screen shot) a frame in some cases.

The problem, I think is coming from the plugs because they are not getting the cache as Alembic and using the default sphere from the standalone, but is doing they on stuff like scatter a custom sphere from the soft which takes a huge time to render cause of the C4D in your case and Maya on our side well not easier I guess is export problem.

In your scene is strange the render time is changing quite a lot one thing to check is the amount of particle on this frames maybe is keeping on the buffer the older particles and get this add on the frames and render really slow ?

Best,
JO

[Bulwerk]

Hellow. Attention! Impossible to render with motion blur!
I wrote 2 times about this problem Aktar. He ignores me.
I'm use Cinema4d. When motion blur is turned on in the scene and there are many particles (hundreds of thousands) the time to calculate motion blur increases frame by frame and can reach huge values ​​in a few hours per frame. BUT! If i restart the render and start from the same frame, the render time is reduced to a few minutes. Because of this, it is absolutely impossible to render projects with a huge number of particles. This error is repeated on any scenes, on any computers. I can’t check this bug in standalone, because it is impossible to export particles from Cinema4d to Stand., but I believe that this problem is related to the core.
Briefly: the same frames with a huge number of particles and turned on MB can be rendered for 5 minutes and for 2 hours.

Please solve the problem !!!

I experience this all the time. I completely avoid concepts and FX that rely on huge amount of particles. Which is a stupid position to have to take in this industry...

[divasoft]

Hellow. Attention! Impossible to render with motion blur!
I wrote 2 times about this problem Aktar. He ignores me.
I'm use Cinema4d. When motion blur is turned on in the scene and there are many particles (hundreds of thousands) the time to calculate motion blur increases frame by frame and can reach huge values ​​in a few hours per frame. BUT! If i restart the render and start from the same frame, the render time is reduced to a few minutes. Because of this, it is absolutely impossible to render projects with a huge number of particles. This error is repeated on any scenes, on any computers. I can’t check this bug in standalone, because it is impossible to export particles from Cinema4d to Stand., but I believe that this problem is related to the core.
Briefly: the same frames with a huge number of particles and turned on MB can be rendered for 5 minutes and for 2 hours.

Please solve the problem !!!

I experience this all the time. I completely avoid concepts and FX that rely on huge amount of particles. Which is a stupid position to have to take in this industry...

This is nonsense! Can a commercial product have such problems? This is not serious.