ComfyUI Node
BSAI VDN-H3 EmptyLatentVideo · H3 Video 5D Empty Latent
H3 video-specific 5D empty latent: [B, 24, T/4, H/16, W/16]. Spatial downsample 16x, temporal downsample 4x, 24-channel video stream. Feeds SamplerCustomAdvanced.latent_image.
BSAI VDN-H3 EmptyLatentVideo · H3 Video 5D Empty Latent
- LATENT
◄width1344►
◄height768►
◄length124►
◄batch_size1►
CategoryBSAI/VDN-H3
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| width | INT | 1344256–4096 | Output video width (pixels); must be a multiple of 16. |
| height | INT | 768256–4096 | Output video height (pixels); must be a multiple of 16. |
| length | INT | 1241–1024 | Output video frame count (temporal downsample 4x; 124 ~= 5s at 24fps). |
| batch_size | INT | 11–16 | Batch size. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| LATENT | LATENT | — |