AI video production and generative film

ComfyUI adds Nvidia RTX nodes for upscaling, interpolation and HDR

Comfy’s 13 October announcement combines local finishing nodes, with hardware and HDR-encoding requirements.

Illustration of a film strip passing a magnifier and light table through successive finishing stages.
AI-generated editorial illustration.

Comfy said on 13 October 2026 that Nvidia RTX video effects now run inside ComfyUI as custom nodes. Its announcement describes a local finishing chain for existing footage: Super Resolution, Frame Generation, then Video HDR.

Super Resolution offers up to 4× scaling or target dimensions up to 8192 pixels, plus same-resolution denoise and deblur modes. Frame Generation interpolates at 2× to 16× or at a chosen timestep. The HDR node converts SDR to BT.2020 HLG or PQ.

The node repository requires an Nvidia RTX GPU and CUDA-enabled PyTorch cu130 or higher; frame generation needs an RTX 40-series GPU or later. CPU-only execution and non-Nvidia GPUs are unsupported.

The README says interpolated footage needs an adjusted encoding frame rate to retain its duration. ComfyUI IMAGE data carries no HDR colour-space metadata, so downstream tools must mark or encode the selected transfer function. The repository is Apache-2.0 licensed; that does not establish the terms for every bundled component. Motion Engineering Today has not tested the nodes.

AI-assisted, reviewed by Codex (15-post review authorized by JJ Rooney, 15 October 2026) on 15 October 2026.