Skip to content

Blendernerf

by Maxime Raafat·Rendering

Images © Maxime Raafat. Used for product review; image policy.

Why we like it

BlenderNeRF earns its place for reducing NeRF and Gaussian Splatting dataset creation to a single-click render-and-camera export directly inside Blender.

What is Blendernerf?

BlenderNeRF is a Blender add-on from Maxime Raafat for building synthetic NeRF and Gaussian Splatting datasets. It outputs renders together with their camera parameters in a single click, and you keep full control over the 3D scene and camera throughout.

Neural Radiance Fields represent a 3D scene as a view-dependent volumetric object from 2D images and their camera information, with the scene reverse engineered from the training images by a simple neural network. Gaussian Splatting is a follow-up method that renders radiance fields in a point-based manner, optimised for GPU rendering and using more traditional graphics techniques to reach high frame rates.

Running NeRF itself is not included with the add-on. With an NVIDIA GPU you can install Instant NGP on your own device, or you can run NeRF in a COLAB notebook on Google GPUs. A fully free and open source version with documentation is available on GitHub, and a demo video is available.

Who is Blendernerf best for?

Blendernerf is best for vFX artists, research fellows and graphics hobbyists creating synthetic NeRF and Gaussian Splatting datasets in Blender.

How much does Blendernerf cost?

Blendernerf is completely free to download and use.

Key features

  • Single-click export of renders and their camera parameters
  • Full user control over the 3D scene and camera
  • Creates synthetic NeRF datasets in Blender
  • Creates Gaussian Splatting datasets
  • Documentation included
  • Fully free and open source version available on GitHub

Tags

#nerf#gaussian splatting#neural radiance fields#dataset#blender#rendering#camera

Frequently asked questions

Do I need an NVIDIA GPU to use BlenderNeRF?

BlenderNeRF handles dataset creation inside Blender, so the plugin itself covers rendering images and exporting camera parameters. Running NeRF is not included with the add-on. If you have an NVIDIA GPU, you can install Instant NGP on your own device by following the instructions on its repository. Without one, you can run NeRF in a COLAB notebook on Google GPUs for free with a Google account. So a local GPU helps but is not strictly required to complete the workflow.

Does BlenderNeRF create Gaussian Splatting datasets as well as NeRF?

BlenderNeRF creates both NeRF and Gaussian Splatting datasets. Neural Radiance Fields represent a 3D scene as a view-dependent volumetric object built from 2D images and their camera information, reverse engineered from the training images with a simple neural network. Gaussian Splatting is a follow-up method that renders radiance fields in a point-based manner, highly optimised for GPU rendering and drawing on more traditional graphics techniques to achieve high frame rates. The add-on produces the renders and camera data both methods rely on.

Is BlenderNeRF free?

BlenderNeRF is listed as free in this catalog. A fully free and open source version with exhaustive documentation is also available on GitHub. For the current licensing terms and any options to support the developer, check the developer's page. The add-on has been developed as a side project across multiple versions of Blender, mainly on macOS, so the developer's page is the best place to confirm what the current release includes.

Can BlenderNeRF use real-world camera footage?

BlenderNeRF is intended for synthetic dataset creation. For real-world footage, separate tools exist for importing motion tracking data from real cameras, such as Tracky by Shopify, an open source iOS app with an adjacent Blender plugin that records motion tracking data from an ARKit session on iPhone. Keep in mind that tracking data can be subject to drifts and inaccuracies, which might affect the resulting NeRF reconstruction quality.

Similar plugins