ComfyUI Node
IronLeo78-MTLTV (v0.0.1, test)
IronLeo78-MTLTV v0.0.1 (test): LTX 2.5 makes the voice, MiniMax H3 makes the picture and lip-syncs to it, and the clip keeps the ORIGINAL LTX voice. Link the video to a Save Video node.
IronLeo78-MTLTV (v0.0.1, test)
- Picture_1
- Picture_2
- Picture_3
- video
- voice
- voice_prompt
- status
◄prompt►
◄voice_prompt►
◄duration5.0►
◄megapixels0.20►
CategoryIronLeo78/MTV
Inputs (7)
| Name | Type | Default | Description |
|---|---|---|---|
| Picture_1 | IMAGE | Main female actress (or the i2v composite). Also the first frame of the LTX voice pass, and it sets the clip's aspect ratio. | |
| prompt | STRING | The MiniMax H3 prompt (the picture), with <Picture 1> .. <Picture 3>. Put the lines in quotes: the voice pass reads this text too when voice prompt is empty. | |
| voice_prompt | STRING | Optional LTX prompt for the SOUND only: who speaks, the lines in quotes, the tone, the room / ambience. Empty = the prompt above without its <Picture N> tags. | |
| duration | FLOAT | 5.01–15 | Seconds, 24 fps, MiniMax frame count (5 s = 124 frames = 5.17 s). At most 15 s. |
| megapixels | FLOAT | 0.200.1–2 | Size of the H3 clip (the voice pass always runs small). Multiples of 32. |
| Picture_2opt | IMAGE | Main male actor (<Picture 2>). | |
| Picture_3opt | IMAGE | Main environment / background (<Picture 3>). |
Outputs (4)
| Name | Type | Description |
|---|---|---|
| video | VIDEO | — |
| voice | AUDIO | — |
| voice_prompt | STRING | — |
| status | STRING | — |