Broadcast Resolution Nodes for ComfyUI

A drop-in custom node suite built for video editors, motion designers, and creators. Generating video with modern AI models shouldn’t require advanced math just to produce a standard frame. Rigid requirements for pixel alignment and megapixel limits can lead to broken latents, CUDA crashes, and unexpected output sizes.

This node pack handles the resolution math automatically, letting you generate clean 4K, 1080p, 720p, and SD video using familiar formats—without worrying about tensor dimensions or sampler errors.

Why It Happens

Modern video diffusion models (LTX, MiniMax, Hunyuan, CogVideo) require frame dimensions strictly divisible by 8, 16, 32, or 64. Because native delivery standards like 1080p or DCI Flat violate these strict macroblock rules, standard samplers either crash outright or output off-spec pixel bounds.



This node suite aims to solve that seamlessly in 2 steps:

1. Input Stage Broadcast Resolution (input)): Computes your true target broadcast or DCI standard, snaps dimensions up to the nearest model-safe multiple so the sampler never crashes, and passes the true target down the pipeline wirelessly.

2. Delivery Stage Broadcast Resolution (output)): Automatically crops, letterboxes, or scales the output video back to the exact target standard right before export.
   

Nodes

1. Broadcast Resolution (input): Selects your target delivery profile and outputs model-safe dimensions for your generator nodes.

2. Broadcast Resolution (output): Native VIDEO standardizer. Automatically syncs with the input profile wirelessly in real time—no target width or height wiring required.

Installation

1. Clone or copy this repository into your ComfyUI/custom_nodes/ directory:

cd ComfyUI/custom_nodes/
git clone https://github.com/UCHIJ/ComfyUI-BroadcastResolutions.git

or install with Comfy CLI

comfy node install comfyui-broadcast-resolutions

2. Restart ComfyUI.

3. Find both the nodes under the Broadcast Video Profile category.