ComfyUI Node
Hy-Nomas Decode
Decodes a world (all its segments, one continuous clip) or a latent into frames.
Hy-Nomas Decode
- decoder
- world
- samples
- script
- frames
- fps
◄seed42►
◄window_overlap5►
◄start_imageone frame►
CategoryHy-Nomas
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| decoder | HYNOMAS_DECODER | — | |
| worldopt | HYNOMAS_WORLD | — | |
| samplesopt | LATENT | Decoded after the world's segments, e.g. the last segment before it is added | |
| scriptopt | HYNOMAS_SCRIPT | Trims to the script's frame count | |
| seedopt | INT | 420–18446744073709550000 | The decoder's noise seed (Tencent's default 42) |
| window_overlapopt | INT | 51–9 | The decoder works on windows of 10 latent frames (40 video frames) that overlap by this many latent frames and are cross-faded. Tencent's 5 decodes every frame twice; 2 is ~1.6x faster with shorter cross-fades. |
| start_imageopt | COMBO | one frame | For a world from Start From Image: keep one frame of the held image (the video opens on it and moves at once), the whole hold, or none. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| frames | IMAGE | — |
| fps | FLOAT | — |