Research

6 posts

lipsync 2.0
research

lipsync 2.0

Introducing lipsync-2.0, the most natural video-to-video lip syncing model. Zero-shot, no training needed — works on live-action, animated, and AI-generated video.

20 Aug 2025 · 2 min read
Lipsync 1.9 beta: a new quality standard
research

Lipsync 1.9 beta: a new quality standard

Introducing lipsync-1.9-beta, the most natural zero-shot lip syncing model. No training data needed — works on live-action, animated, and AI-generated video.

15 Jan 2025 · 2 min read
Lipsync-1.8.0-beta in research preview
research

Lipsync-1.8.0-beta in research preview

Announcing lipsync-1.8.0-beta, a new zero-shot lip sync model that works on any video in the wild with significantly improved teeth generation and visual quality.

10 Oct 2024 · 1 min read
Audio-Visual Face Reenactment
research

Audio-Visual Face Reenactment

A novel method for generating realistic talking head videos using audio and visual streams, with improved lip sync quality and identity-aware generation.

6 Oct 2022 · 1 min read
A lip sync expert is all you need
research

A lip sync expert is all you need

Wav2Lip introduces a powerful lip-sync discriminator for accurate speech-to-lip generation on arbitrary identities in unconstrained talking face videos.

23 Aug 2020 · 1 min read
Speaking styles for lip-to-speech synthesis
research

Speaking styles for lip-to-speech synthesis

Exploring lip-to-speech synthesis by learning individual speaking styles. Includes a large-scale benchmark dataset and a novel approach for natural speech generation from lip movements.

17 May 2020 · 1 min read
Back to Blog