Cinema 4D plug-in renders slow than expected..

[professorhaddock]

Hey guys,

I just purchased a Cinema 4D plug-in / standalone license yesterday and have been playing with actually sending renders to the picture viewer, and I'm disappointed. Truth be told, I'm limited to version 2.17 because I'm still on Mountain Lion, but I'm wondering whether that'd really speed things up. Anyway, a scene that'd take 5 to 6 minutes to render with C4D's physical render (with GI and AO), takes 8 to 10 minutes with Octane. I thought the whole idea behind Octane was the tremendous speed gains afforded by the GPU. And curiously, the Octane Live Viewer cleans up fairly rapidly (I only have it set to around 700 samples), but then the "preparation" stage of the actual render takes forever.

I'd appreciate any suggestions.

[aoktar]

Welcome,
propably doing lot of mistakes about optimisations of scene. You can't compare PR with Octane's speed. Would be great to know your gpu? Have you watched the our tutorials?

[professorhaddock]

What optimizations are you talking about specifically?

And in my signature it says I have a 770GTX with 4GB. And yes, I have watched the tutorials, but what ones are you talking about specifically?

[aoktar]

1- Lot of geometry will cause long time to send and compile for gpu. That's the reason why we have a SMART ENGINE. It just updates the changed parts while you are working or making final animations. If you use it very well then will be much shorter than any other engine's preparations times. Also you should tweak your meshes to use less polygons. Ex. set render subdivisions of hypernutbs, etc.. to generate less triangles.

2- Set samples low as possible.

3- See Inlifethrill's render optimisations and quick start video.

4- Demo versions is for testing better to do purchases. It may not fit your setups or requests. But It's impossible to hit the performance by a cpu renderer especially by PR. Propably you're not tweaking well. Or you haven't tested for all kind of scenes. If you post some scenes so someone can help to you for tweaks.

[whersmy]

Haddock,

a gtx770 scores even less than a gtx580 when it comes to speed. It can be fast, but still it`s more or less an entry level card.
Are you rendering large res? Lots of specular materials? Many mesh emitters? Pathtracing?

[professorhaddock]

aoktar:

1. I'm well aware of using less geometry to be more efficient, I've been doing this for a while.

2. Yeah

3. I saw it

4. The demo version doesn't give you an accurate picture of how it'll perform because it limits the size of the final render. According to the live viewer, everything was just fine. Funny how I could think it'd render fine, no?


whersmy:

I'm well aware it's not anywhere near the top of the heap, but calling it "entry level" is a little far. I'm rendering at 720p, not too many specular materials, and I'm using DirectLighting.

[aoktar]

aoktar:
4. The demo version doesn't give you an accurate picture of how it'll perform because it limits the size of the final render. According to the live viewer, everything was just fine. Funny how I could think it'd render fine, no?

1000×600 that's the limit of output. So you can set both renderer to this size and you would compare easily. So that's not logical that to say impossible. I'm still curious about what your comparisions. Any images, scenes to illustrate problem?

[professorhaddock]

I was talking about using 720p as a comparison point.

I don't have any images or video at the moment; the problem is simply it's taking more time than expected. The "preparing" phase of the render is what's taking so long. Does it only need to "prepare" once for an entire animation sequence or does it need to happen each frame?

[jayroth]

Start turning things off in Cinema to see where the issue lies. Hair can be one of the bottlenecks; the more hairs you have, the longer it will take for Octane to process the scene. Dynamics can slow things down also; cache them (true for any render engine). Convert cloners and similar items to minimize the regeneration of that portion of the import. You'll soon discover where you're bottlenecks are, and then proceed from there...