ComfyUI Node
WanMaskToLatentSpace
A ComfyUI node in WanVaceAdvanced with 8 inputs and 1 output.
WanMaskToLatentSpace
- mask
- mask
◄spatial_methodmax►
◄temporal_methodmax►
◄first_frame_specialtrue►
◄target_width0►
◄target_height0►
◄expand_spatial0►
◄expand_temporal0►
CategoryWanVaceAdvanced
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| mask | MASK | — | |
| spatial_method | COMBO | max | Spatial reduction: max=preserve any coverage, min=require full coverage, mean=average, area=area-based, nearest=nearest-exact |
| temporal_method | COMBO | max | Temporal reduction: max/min/mean across 4-frame groups, or first/last frame only |
| first_frame_specialopt | BOOLEAN | true | True: first frame 1:1, rest grouped by 4. False: all frames grouped by 4 uniformly. |
| target_widthopt | INT | 00–16384 | Target latent width. 0 = auto (mask width / 8) |
| target_heightopt | INT | 00–16384 | Target latent height. 0 = auto (mask height / 8) |
| expand_spatialopt | INT | 0-16384–16384 | Pixels to grow (+) or shrink (-) spatially before downscaling |
| expand_temporalopt | INT | 0-1000–1000 | Frames to grow (+) or shrink (-) temporally before downscaling |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| mask | MASK | — |