Get help
Documentation, Support and Frequently asked questions
Documentation:
Frequently
asked questions
Search topics and steps to find the right answers in fewer clicks.
What is VDB compression and how does it work in Zibra AI?
ZibraVDB compresses OpenVDB effects, allowing real-time rendering in game engines by reducing data size up to 20 times. This enables high-quality 3D visuals in Unreal Engine 5.
How is ZibraVDB different from NeuralVDB or NanoVDB?
The goal of ZibraVDB is not just to reduce size, but to enable real-time rendering of large VDB sequences. This is possible because of our real-time decompressor and custom volumetric renderer optimized specifically for this type of data. With ZibraVDB, you can load the entire compressed sequence into memory and decompress the needed frame on demand. This is much faster than loading each frame from the disk.
How is ZibraVDB different from Unreal’s Heterogeneous Volumes?
Foremost, ZibraVDB is twice as fast as Heterogeneous Volumes in terms of rendering and importing speed. Additionally, ZibraVDB allows you to render large sequences, like 80GB, in real-time, while Unreal Engine crashes even when importing a 10GB sequence. ZibraVDB is also much easier to use; you simply drag and drop your effect into the scene, and it works with no additional setup needed!
ZibraVDB also integrates with Heterogeneous Volumes for those who still require it. You’ll benefit from a faster importing process and support for large effects. However, the rendering speed will be much slower compared to using the custom ZibraVDB renderer.
How much the size of the VDB sequence can be reduced?
This depends a lot on the effect itself and how much detail you want to preserve. On average, you can get 40x compression without noticeable quality loss. With our upcoming temporal compression, you’ll be able to reach a 150x compression rate.
What maximum size of the VDB sequence can I use in real-time?
This depends on the size of your largest frame in the sequence. After decompressing this frame, it should fit into your VRAM while still leaving some space for the compressed sequence and rendering resources. You can even compress and render a 1TB effect if it has many frames that aren't too large. ZibraVDB also makes additional optimizations for single-frame VDB (e.g., clouds) or single-channel sequences, allowing you to render frames larger than 2GB. Be aware, though, that rendering frames with billions of voxels can be quite expensive.
How much realism can I get with ZibraVDB?
ZibraVDB renderer is designed for maximum speed. It also has advanced features like projected shadows, precise reflections, and can support illumination from up to 15 different lights. Realism largely depends on how you direct your effects and integrate them into the scene. Note that we don’t have light scattering for now. It’s better to bake all emission in the volume to the temperature channel if you want to import your complex shaded effects from Houdini. ZibraVDB also integrates with Heterogeneous Volumes, which is useful if you want to use Path Tracing with ZibraVDB for maximum quality.
Does ZibraVDB support the compression of particle cache stored in VDB format?
No, VDB is not the best way to store particle data efficiently. However, we do plan to create a solution for compressing particles and geometry.
Contact support