# YuanxunLu/LiveSpeechPortraits

Live Speech Portraits: Real-Time Photorealistic Talking-Head Animation (SIGGRAPH Asia 2021)

Repository: https://github.com/YuanxunLu/LiveSpeechPortraits
Canonical: https://ross.abutalabs.com/products/livespeechportraits
Language: Python
License: MIT
License Family: permissive
Last push: 2023-06-19T21:51:32+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1854, "days_push": 1171, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1283, forks 216 (observed 2026-08-28T04:04:14.481622+00:00)

## What it is
A PyTorch implementation of the SIGGRAPH Asia 2021 paper 'Live Speech Portraits', which generates photorealistic personalized talking-head animation from audio input in real time (30+ fps). It combines audio feature extraction, head pose and upper-body motion prediction, and image-to-image translation to synthesize high-fidelity facial renderings.

## Use cases
- generate a talking-head video from an audio clip
- animate a portrait photo driven by speech audio
- create real-time photorealistic avatars from voice
- research on audio-driven facial animation
- synthesize head poses and facial dynamics from speech
- reproduce SIGGRAPH Asia 2021 talking-head paper results

## When to choose
- you need audio-driven photorealistic talking-head generation with a specific trained person model
- you want a research baseline or to build on a published real-time talking-head method
- you have a GPU environment and can work with research-grade code

## When to avoid
- you need a production-ready product with a polished UI or API
- you need to animate arbitrary new identities without training data for that person
- you need actively maintained code with recent dependency support
- you need full-body or non-portrait avatar animation

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, video-processing, audio-processing, computer-vision, speech-recognition
- domain: computer-vision, deep-learning, artificial-intelligence, media
- platform: python, windows
- tags: talking-head, audio-driven-animation, photorealistic, face-synthesis, research-code, siggraph, video, audio, linux, gpu

## Member repositories
- YuanxunLu/LiveSpeechPortraits (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:14.481622+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:56:47.668328+00:00, confidence not recorded.
  - readme: https://github.com/YuanxunLu/LiveSpeechPortraits (fetched 2026-08-28T04:04:14.481622+00:00, sha e796019e21af)
- Data as of 2026-08-30T08:39:29.467469+00:00.