Rendering Heavy OpenVDB Volumes in Octane: Step Length, Density and VRAM
The settings that decide how a VDB looks in Octane and how long it takes to render:
- Step length first: Volume Step Length sets how often Octane samples the volume along a ray. The default is 4 m; a smaller step gives more detail and renders more slowly (OTOY).
- Then density: OTOY recommends setting the step length to an acceptable value first and reducing density after, or the volume can render like a solid object.
- Memory per GPU: every GPU holds the scene, VDB grids included, in its own VRAM, so trim unused grids before you render on several cards.
- On iRender: 1x to 8x RTX 4090 (24 GB) or RTX 5090 (32 GB) servers with Cinema 4D and OctaneRender pre-installed, and a local NVMe drive for VDB sequences.
Smoke, fire and clouds from Houdini, EmberGen or Cinema 4D Pyro usually reach Octane as OpenVDB files. They can look washed out, blocky or solid, and they can render slowly. Most of that comes down to three things: the volume step length, the density and absorption balance, and how much data each frame of the VDB sequence carries. This guide goes through each with OTOY’s documentation and ends with a way to test a heavy volume on your own hardware.
1. How Octane renders a VDB
In Cinema 4D you add a volume from Live Viewer › Objects › Add Octane VDB Volume (or Add Octane Fog Volume for generated volumes), and point it at a single VDB file or an animated sequence. Every volume object has a Medium tab that controls how light behaves inside it: density, step length, absorption, scattering and emission (OTOY: Volume Mediums (Cinema 4D)). Octane renders volumes in an unbiased way, so light can scatter several times inside them (OTOY: Volumes (Houdini)), which is what makes dense smoke look soft and also what makes it expensive.
2. Volume Step Length: detail against render time
Octane ray-marches through a volume: it takes steps along each ray and samples the volume at each step. Volume Step Length is the distance between those steps. OTOY’s documentation gives the rules (OTOY: Volumetrics (Cinema 4D)):
- The default step length is 4 m. If your volume is smaller than that, decrease it.
- If the step length is far larger than the volume, the ray takes a single step through the whole volume, and the detail disappears.
- The smallest step gives the most accurate result, and a smaller step renders more slowly.
- Set the step length to an acceptable value first, then adjust the other settings.
Scene scale matters here: a 4 m step is meaningless for a 30 cm candle flame. Check your scene units and the size of the VDB before judging the look. In the newer Standard Volume Medium, the shadow step is set as a percentage of the voxel size instead of an absolute distance, which ties the step to the resolution of the simulation (OTOY: Standard Volume Medium).
3. Density, absorption and scattering
Density multiplies the scattering of the volume, so a high density with a large step can look like a solid object. OTOY’s advice is to reduce the step length to an acceptable level of performance and accuracy first, and reduce density after (OTOY: Volume Mediums (Standalone)). Then set the colours:
- Absorption darkens the volume and controls how much light passes through.
- Scattering brightens it and gives smoke its soft, lit look.
- Emission, usually driven by a temperature or flame grid through a gradient, makes fire glow. Cinema 4D’s gradient presets are a good start for fire.
Change one at a time and judge the result in a full render, because the Live Viewer converges more slowly in dense volumes.
4. Keep the VDB light enough for every GPU
A VDB file can contain several grids: density, temperature, flame, velocity and others. Octane only needs the grids you actually connect to the medium, but large files still cost loading time and memory. Before rendering a heavy sequence:
- Export only the grids you render. Velocity is useful for motion blur; otherwise leave it out, together with simulation-only fields.
- Match the voxel size to the shot. A simulation run at a much finer resolution than the camera can resolve adds memory without adding visible detail.
- Keep sequences on a local SSD. An animated VDB is read from disk every frame; reading it from a network share adds wait time to every frame.
On several GPUs, each card holds the scene in its own VRAM; Octane does not combine memory across cards. Octane can keep geometry and textures out of core in system RAM, which is slower than VRAM (OTOY: Out of Core). So the VRAM of one card is the limit to plan for: 24 GB on an RTX 4090 and 32 GB on an RTX 5090.
5. Measure a heavy volume before the full render
We have not published timed tests of volume settings, and the effect of a step length or density change depends on the simulation, so this guide does not quote speed-ups. Measure on your own shot:
- Pick the frame where the volume is largest on screen.
- Render it at final samples and note the time and the VRAM use in Octane’s device information.
- Halve the step length and render again; compare detail and time.
- Keep the largest step length that still holds the detail you need, then multiply the time per frame by the frame count and check the price page.
For reference, the Octane runs iRender has recorded on its servers are on the multi-GPU render benchmarks page.
SaaS or IaaS render farm for heavy VDB volumes?
Volume settings are saved in the scene, so a VDB renders the same on any farm that runs the same Cinema 4D and Octane versions, as long as the VDB files travel with the project. The difference is where you can tune them. On a SaaS render farm you upload the finished scene and the farm returns frames. On an IaaS render farm such as iRender you rent a GPU server, open the scene yourself and adjust step length and density on the render hardware.
| Your situation | SaaS render farm | IaaS render farm (iRender) |
|---|---|---|
| The look is approved and the VDBs are final; you need frames | A good fit: upload, submit and download frames, with no machine to manage | Works too, but you start the server and run Cinema 4D yourself |
| You have a deadline and need to know when rendering starts and ends | Jobs join the farm’s queue, so the start time depends on how busy the farm is; many farms sell higher-priority plans to start sooner | The server runs only your job from the moment it boots, so you control when rendering starts and can time the whole job |
| You are still tuning step length and density | You tune at home, then submit | Render test frames at final quality on an RTX 5090 and compare |
| The VDB sequence is large | Check how the farm collects and stores VDB sequences before you submit | Upload the sequence once to iRender Drive and copy it to the server’s local NVMe drive |
| The scene may not fit in 24 GB | The log tells you after the job | Check VRAM use on a 32 GB card before committing |
| A deadline with many frames | The farm spreads frames across its machines | Rent a 4x or 8x server, or several servers with split frame ranges |
A comparison of how the two farm models work for this task, not a speed test.
A simple rule: if the volume is approved and the files are final, a SaaS farm is the shorter path. If you are still tuning the look or checking memory, you need a machine you can open and measure, which is what an IaaS server gives you.
A volume test workflow on an iRender server
- Upload first. Put the scene and VDB sequence in iRender Drive before you start a server; transfer and storage are free and no rental time runs while you upload.
- Start a server. Cinema 4D and OctaneRender are pre-installed; sign in with your OTOY account. Octane uses every GPU in the server, so a 4x or 8x server suits final renders; multi-GPU servers bill per minute.
- Copy the VDBs to the local drive and relink them in the scene.
- Tune and measure on the heaviest frame as in section 5.
- Render the sequence, or split frame ranges across several servers for a deadline.
Frequently asked questions
1. Why does my VDB look blocky or washed out in Octane?
Usually the Volume Step Length is too large for the size of the volume. The default is 4 m; decrease it for small volumes, then adjust density.
2. What Volume Step Length should I use in Octane?
As small as gives the detail you need at an acceptable render time. Start below the size of the volume, render a test frame, then halve it and compare.
3. Why does my Octane volume render like a solid object?
The density is too high for the step length. OTOY recommends setting the step length first and then reducing the density.
4. Does Octane combine VRAM across GPUs for volumes?
No. Each GPU holds the scene in its own VRAM, so plan for the memory of one card: 24 GB on an RTX 4090 or 32 GB on an RTX 5090.
5. How can I make VDB sequences render faster?
Export only the grids you use, match the voxel size to the shot, keep the files on a local SSD, and use the largest step length that keeps the detail you need.
6. Can I use VDB files from Houdini or EmberGen in Cinema 4D Octane?
Yes. Load them in an Octane VDB Volume object and connect the grids, such as density and temperature, in the medium.
Render your Cinema 4D and Octane projects on iRender
iRender is an IaaS GPU render farm: you rent a bare-metal server with 1, 2, 4 or 8 NVIDIA RTX 4090 (24 GB) or RTX 5090 (32 GB) GPUs, an AMD Ryzen™ Threadripper™ PRO CPU, 256 GB ECC RAM and a 2 TB NVMe SSD, connect by remote desktop and work on it like your own workstation, with full administrator access. Your server image is saved between sessions, so your plugins and settings stay installed. Cinema 4D and OctaneRender are pre-installed; sign in with your own OTOY account.
Multi-GPU servers are billed per minute (1-GPU servers have a one-hour minimum), and sessions of 3 hours or more get 10% off. Our 4-GPU and 8-GPU servers are liquid-cooled; in our own measurements the GPUs hold 60–70°C under sustained render load. Elite support is available 7 days a week by hotline, live chat, Skype, WhatsApp and email.
This quick start shows how to start a server and render Cinema 4D and Octane on it:
Two Octane runs we recorded on our servers: 300 frames of a car drift scene at 1080p on 4x RTX 5090 in 1h 18m (15.6 seconds per frame), and the same scene on 4x RTX 4090 at 18.8 seconds per frame:
New users get a 100% bonus on the first top-up (minimum $50). See plans on the Octane render farm page or download the iRender GPU app.
