RuneXX

/

LTX-2.3-Workflows

like 615

i was thinking to try that out, for a short music video.. or could be any audio really.
Will see if i can make that work well ;-)

Available here:
https://huggingface.co/RuneXX/LTX-2.3-Workflows/tree/main/Movie-Maker/Prompt-Relay-Custom-Audio
That is the one with single input image, prompt relay + custom audio.

I'll upload a version with multi input images as well

thanks - you are like the new kajai - these workflows are awesome

Happy you like ;-) credits to the real Kijai though... hehe.. he is the one that made the nodes, and did the harder part ;-)

kindly ask you to add a "custom audio" switch to add a mode without mp3. Or make a separate workflow without mp3. This wonderful scheme with the addition of prompts and screenshots to each scene is very good, but in this version it only works for songs; I want to make separate scenes with dialogues. Thank you. Its about workflow: LTX-2.3_-_ I2V_T2V_Music-Video-Creator_multi-scene_custom_audio

kindly ask you to add a "custom audio" switch to add a mode without mp3. Or make a separate workflow without mp3. This wonderful scheme with the addition of prompts and screenshots to each scene is very good, but in this version it only works for songs; I want to make separate scenes with dialogues. Thank you. Its about workflow: LTX-2.3_-_ I2V_T2V_Music-Video-Creator_multi-scene_custom_audio

How would you do the audio part? Prompt it? If so, you can use the non custom audio one here: https://huggingface.co/RuneXX/LTX-2.3-Workflows/tree/main/Movie-Maker/
Or do you mean using something like a TTS?

No, I meant the usual workflow with sound from LTX, without MP3 substitution at all. As done here: LTX-2.3_-_I2V_T2V_Basic. So that can substitute a prompt and a screen for each segment. Then LTX will voice the entire scene herself. Well, or add the ability to add own MP3 to each segment. Then there will be a chance not only for 1 song,
but also for entire scenes with dialogues. Unfortunately, I couldn’t turn on Movie-Maker...
Thank you

UPD: everything is fine. What I wanted in the "LTX-2.3_-_I2V_Short-Story_PromptRelay-Timeline_multi-image_multi-sequence scheme"
Great, thank you very much :) You are the best, wonderful worrkflows :) .

Most welcome ;-)

using the custom audio workflow - would it be possible to have 2 speakers i.e. upload 2 audio and one person says the first audio and the other one says the 2nd audio ?

using the custom audio workflow - would it be possible to have 2 speakers i.e. upload 2 audio and one person says the first audio and the other one says the 2nd audio ?

LTX only takes one audio reference input, so uploading 2 audio files wouldn't work directly, but there are ways

• Merge your 2 audio files into 1 audio file, so that the 1 audio file has the dialog between both persons
  (this works in the workflow already, just prompt accordingly .. like "First the man speaks. Then the woman speaks" etc)

• Use a TTS node that supports multi-speaker (FishAudio, Omni Audio etc). And then in the TTS node add both your mp3 audio files, one per speaker
  (the tts then clones the voices, both of them, and you prompt what they should say, and the TTS node feed the combined dialog to LTX)

The 2nd method would be a combo of this workflow, and the TTS workflows here https://huggingface.co/RuneXX/LTX-2.3-Workflows/tree/main/Talking-Avatar-TTS

But the idea of having 2 speakers is a good one, might add that to the workflow in some way (maybe some clever audio merge node exists). And I can make a TTS version for sure.

Could also use ID Lora, and have 2 references.
But this would then be one speaker per scene (they can both be visible, but one talking). But that can work quite well. Rarely do both people talk at same time, if ever ;-)
And then switch who is talking per scene