Run NVIDIA DLSS 5 Neural Rendering over your frames and videos, inside ComfyUI.
DLSS 5 is NVIDIA's neural rendering pass. In a game it reconstructs the material and lighting detail that real-time rendering has to leave out, such as skin subsurface scattering, light transmission through hair and contact shadows, guided by the engine's own buffers. This node pack drives that same renderer over video frames and image batches, with optional 1.5x to 3x upscaling. Without those buffers what it recovers here is material response: skin, hair and fabric structure.
It is not a filter that imitates the look. Frames go to the native renderer and come back reconstructed.
- How it works
- Requirements
- Installation
- Nodes
- Recommended settings
- Example workflows
- Verification
- Performance
- Known limitations
- Troubleshooting
- Project layout
- Licensing and attribution
In games, DLSS 5 runs as "3D-Guided" Neural Rendering: the engine hands the model its rendered frame together with geometry, texture and lighting buffers, and motion vectors. Video has none of that. What reaches the model here is the decoded frame plus motion vectors estimated from the footage itself, so the guidance is weaker than in a game, and the reconstruction is correspondingly more conservative.
There is no Python API for any of it. The renderer lives in a native D3D12 process that hosts ReShade and the RenoDX DLSS 5 add-on and evaluates NGX feature 18. This node pack implements the client side of that worker's binary protocol:
ComfyUI node
|
| RGBA8 frame + FP16 motion vectors + history reset flag
v
native D3D12 worker
(ReShade carrier -> RenoDX DLSS 5 add-on -> NGX feature 18)
|
| RGBA8 reconstructed frame
v
ComfyUI node
Encoded video carries no motion vectors, so they are estimated per frame with dense optical flow (OpenCV DIS) and supplied at the render resolution, including automatic history resets on scene cuts. Everything else is handled here: session setup, dimension negotiation, frame ordering, cancellation and diagnostics.