screencastgen

Lipsync Facade

Facade for lip-sync video generation from audio + reference video.

Source: screencastgen/lipsync.py

Overview

Provides a stable API for lip-sync generation, delegating to providers in Lipsync Registry. Handles video looping to match audio duration.

Functions

`generate_lipsync_video(audio_path, reference_video_path, output_path, provider, device, latentsync_preset) -> str`

Generates a lip-synced video from audio and a reference face video.

Parameters:

Param	Type	Default	Description
`audio_path`	`str`	—	Path to audio file
`reference_video_path`	`str`	—	Path to face reference video
`output_path`	`str`	—	Output video path
`provider`	`str`	`"auto"`	Lipsync provider name
`device`	`str`	`"auto"`	Compute device
`latentsync_preset`	`str`	`"quality"`	LatentSync quality preset

`_get_audio_duration(audio_path) -> float`

Returns duration of an audio file in seconds.

`_loop_video_to_duration(video_path, duration, output_path) -> str`

Loops a reference video to match the target audio duration using ffmpeg.

Dependencies

Lipsync Facade
├── Lipsync Registry       (provider dispatch)
├── TTS Base               (resolve_device)
├── ffmpeg CLI                 (video looping)
└──▶ consumed by Lipsync Pipeline
     └──▶ or Remote GPU Client (remote mode)

screencastgen

Lipsync Facade

Overview

Functions

generate_lipsync_video(audio_path, reference_video_path, output_path, provider, device, latentsync_preset) -> str

_get_audio_duration(audio_path) -> float

_loop_video_to_duration(video_path, duration, output_path) -> str

Dependencies

See Also

`generate_lipsync_video(audio_path, reference_video_path, output_path, provider, device, latentsync_preset) -> str`

`_get_audio_duration(audio_path) -> float`

`_loop_video_to_duration(video_path, duration, output_path) -> str`