ComfyUI Node
H3 Context Taper Noise (latent)
Inject tapered gaussian noise into the tail latent steps of an H3 AV context latent, for H3 Motion Context's context_latent input. Audio passes through untouched. Output is for context_latent only -- do not wire it into a VAE decode.
H3 Context Taper Noise (latent)
- context_latent
- context_latent
- alpha_schedule
◄tail_frames22►
◄alpha0.45►
◄alpha_end0.10►
◄ramp_steps2►
◄seed0►
◄noise_scalematch_latent_std►
Categoryconditioning/minimax
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| context_latent | LATENT | The previous clip's saved AV latent, straight from H3 Motion Context Load Latent. | |
| tail_frames | COMBO | 22 | Pixel frames to inject, converted to latent steps. Only these lengths are whole numbers of steps: 5/22/39/56 -> 2/7/12/17 steps. Set this to the same value as Motion Context's context_length. |
| alpha | FLOAT | 0.450–1 | Flat injection strength across the pinned steps. |
| alpha_end | FLOAT | 0.100–1 | Strength at the last latent step, the one touching the seam. |
| ramp_steps | INT | 21–64 | Latent steps at the end that taper down to alpha_end. Steps are coarse: on a 22-frame window one step is 4 pixel frames, so 2 is roughly the pixel recipe's 3-frame ramp. |
| seed | INT | 00–18446744073709550000 | — |
| noise_scaleopt | COMBO | match_latent_std | match_latent_std scales the gaussian to the carried window's own standard deviation, so alpha means the same thing whatever the latent's scale. unit uses raw N(0,1). |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| context_latent | LATENT | — |
| alpha_schedule | STRING | — |