How to use custom voice AND face with Seedance 2? step by step workflow. bookmark this thread 🧵

tools we'll use: * seedance 2: * voice-cloner: both has to be on mitte. other platforms (runway etc) don't support audio properly.

step 1. clone your voice. entirely for education purposes, I'll clone @xueqinjiang's voice. I love him and it'd be a great demo ❤️ so open voice cloner: and upload a voice sample (45 seconds), type name + description, and choose an avatar

step 2. generate speech once the voice is cloned, mitte places your new voice character under "Voice Library" tab in the model menu. open that and type any script in your new voice model. you can optionally choose emotion and speed. it produces incredibly realistic result

step 3. create the video you have the audio now, and video is next. select "seedance 2" again from the model menu. add your character's image first. then right click in the audio you made and click "add to prompt" enter the prompt like in the image below

step 4. done make sure your prompt tags the audio properly. my prompt was "lecture by the teacher in @image1 speaks the given @audio1" make sure your video length matches the audio. enjoy the result 🥳

Great tutorial, Ink! Thanks for sharing!

you’re welcome 🫶

illegal, it's people like you that give ai art a bad name

Mitte is great, hope they will release an API soon to enable automation

hmmm, i see voice cloner. Btw, if i make a video, should i make them mute then inster the voice? or use the voice as reference? Which one should i do first?

create the voice first, then add it to the video. as i explain here

oh ok, should i match each second of the word with his mouth or just let it read from previous video? I sticthed several videos into a video.

model does the lipsync for you

oh okayy. i will try when i need it. Thanks Ink!

@brubbleR stick with your ltx and localhost then lol

Great 😸

ty