Workflow: V2V Just Dub It - lip synced multi-language dubbing with IC-Lora-LipDub
Italian
Swedish
German
Spanish
V2V Just Dub It - lip synced multi-language dubbing with IC-Lora-LipDub
Translate any video with LTX official LipDub lora, based on the JustDubIt paper.
Lora available here: https://huggingface.co/Lightricks/LTX-2.3-22b-IC-LoRA-LipDub
And workflow to try here: https://huggingface.co/RuneXX/LTX-2.3-Workflows/tree/main/Video-2-Video
You need the very latest of ComfyUI-LTXVideo. (update in comfyui manager, or install if you dont have already)
you mistook spanish for german and german for spanish...so funny
I'm confused on the differences between JustTalk, JustDubIt, and now LipDub. Is LipDub just the latest (and thus superior) of the three?
Regardless, the videos look super nice and this seems great.
Not sure about JustTalk (i can't find any info from google, was it MetaClaw?), but there is this notice at Just-Dub-It repository:
Notice: This repo is now archived, please move to our offical repository and checkout the latest LTX2.3 support on lipdub.
So i guess LipDub is a (official) continuation of JustDubIt 🤔
you mistook spanish for german and german for spanish
ups.. haha.. fixed ;-)
So i guess LipDub is a (official) continuation of JustDubIt 🤔
Seems like it ;-) Its based on the same paper
I'm confused on the differences between JustTalk, JustDubIt, and now LipDub
JustDubIt = LipDub
It was made by someone else for LTX-2.0 and there was a workflow for that back then. For LTX-2.3, LTX themselves made a lora based on the same paper.
I used the same name for the wf as in LTX-2.0.
JustTalk (with dubbing)
Its a bit different. .Its for adding sound to any silent video, mostly. And made before LTX made the LipDub lora.
It uses a mask over the mouth area, and inpaints. But can be used much the same way also.
Probably will remove the wf dub version, since its not needed anymore...
Btw LipDub doesn't use mask, right? since it can changed the facial motion too, instead of just the lips 🤔
LipDub lets you take an existing video and change what's being said, with regenerated speech and lip motion that stay aligned naturally in the final video.
Instead of patching lips after the fact, the model generates voice and facial motion together — making dubbing more stable across motion, angles, and real-world footage.
Btw LipDub doesn't use mask, right? since it can changed the facial motion too, instead of just the lips 🤔
No masking for LipDub, it uses a lora