1.LatentSync: Taming Audio-Conditioned Latent Diffusion Models for Lip Sync with SyncNet Supervision 字节 2024

文章地址：https://arxiv.org/pdf/2412.09262

代码地址：https://github.com/bytedance/LatentSync    训练推理都有

2.wan2.2-s2v 阿里通义 20250826

文章：[2508.18621] Wan-S2V: Audio-Driven Cinematic Video Generation

代码：https://github.com/Wan-Video/Wan2.2 只有推理

3.Let Them Talk: Audio-Driven Multi-Person Conversational Video Generation 中山大学an美团20250528

文章：https://arxiv.org/pdf/2505.22647

代码：https://github.com/MeiGen-AI/MultiTalk 只有推理

4.Loopy: Taming Audio-Driven Portrait Avatar with Long-Term Motion Dependency

字节and浙大 支持Singing

文章：https://arxiv.org/pdf/2409.02634   ICLR2025

代码：只有demo Loopy: Taming Audio-Driven Portrait Avatar with Long-Term Motion Dependency

5.EchoMimic: Lifelike Audio-Driven Portrait Animations through Editable Landmark Conditioning

AAAI 2025 20240711 蚂蚁 Pretrained models with better sing performance to be released

项目EchoMimic: Lifelike Audio-Driven Portrait Animations

文章[2407.08136] EchoMimic: Lifelike Audio-Driven Portrait Animations through Editable Landmark Conditions

代码https://github.com/antgroup/echomimic 只有推理

EchoMimicV2: Towards Striking, Simplified, and Semi-Body Human Animation. GitHub

20250227 CVPR 2025

EchoMimicV3: 1.3B Parameters are All You Need for Unified Multi-Modal and Multi-Task Human Animation. GitHub  20250708

6.EMO: Emote Portrait Alive - Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions 20250227 阿里 ECCV2024

项目EMO

文章：[2402.17485] EMO: Emote Portrait Alive -- Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions

代码：没内容GitHub - HumanAIGC/EMO: Emote Portrait Alive: Generating Expressive Portrait Videos with Audio2Video Diffusion Model under Weak Conditions

EMO2: End-Effector Guided Audio-Driven Avatar Video Generation 20250118阿里

项目：EMO2。支持Singing

文章：[2501.10687] EMO2: End-Effector Guided Audio-Driven Avatar Video Generation

7.VASA-1: Lifelike Audio-Driven Talking Faces Generated in Real Time

Microsoft Research Asia 20240416 NeurIPS 2024 (Oral) 

项目：https://www.microsoft.com/en-us/research/project/vasa-1/

文章：[2404.10667] VASA-1: Lifelike Audio-Driven Talking Faces Generated in Real Time

8.FaceFormer: Speech-Driven 3D Facial Animation with Transformers, CVPR 2022.

文章：https://arxiv.org/pdf/2112.05329

代码：https://github.com/EvelynFan/FaceFormer?tab=readme-ov-file 有训练代码

9.SkyReels-Audio: Omni Audio-Conditioned Talking Portraits in Video Diffusion Transformers MimicMotion    SkyReels Team, Skywork AI 20250601 支持唱歌

项目SkyReels-Audio

文章https://arxiv.org/pdf/2506.00830

代码https://skyworkai.github.io/skyreels-audio.github.io/ 仅推理

SkyReels-A1: Expressive Portrait Animation in Video Diffusion Transformers

SkyReels-A2: Compose Anything in Video Diffusion Transformers

SkyReels-A3：Towards Ultra-Long Audio-Conditioned Video Generation

10.InfiniteTalk: Audio-driven Video Generation for Sparse-Frame Video Dubbing

20250819 多家单位

项目InfiniteTalk

文章[2508.14033] InfiniteTalk: Audio-driven Video Generation for Sparse-Frame Video Dubbing

代码：只有推理GitHub - MeiGen-AI/InfiniteTalk: ​​Unlimited-length talking video generation​​ that supports image-to-video and video-to-video generation

11.MusicFace: Music-driven expressive singing face synthesis

20240201 厦大 没开源

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10897677&tag=1[2508.14033] InfiniteTalk: Audio-driven Video Generation for Sparse-Frame Video Dubbing

12.FantasyTalking: Realistic Talking Portrait Generation via Coherent Motion Synthesis

阿里 MM2025

文章https://arxiv.org/pdf/2504.04842

代码 https://github.com/Fantasy-AMAP/fantasy-talking只有推理

13.HHunyuanVideo-Avatar: High-Fidelity Audio-Driven Human Animation for Multiple Characters腾讯混元 20250603

文章https://arxiv.org/pdf/2505.20156

代码https://github.com/Tencent-Hunyuan/HunyuanVideo-Avatar 只有推理

14.DiffSynth-Studio

开源项目GitHub - modelscope/DiffSynth-Studio: Enjoy the magic of Diffusion models!

15.SadTalker: Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation  

CVPR2023

文章[2211.12194] SadTalker: Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation

代码 仅推理GitHub - OpenTalker/SadTalker: [CVPR 2023] SadTalker：Learning Realistic 3D Motion Coefficients for Stylized Audio-Driven Single Image Talking Face Animation

16.Speech2Vid

17.Wav2Lip

18.DeepFaceLive

19.Easy-Wav2

20.VideoReTalking

21.UniTalker: Conversational Speech-Visual Synthesis

20250807 MM2025

文章

代码https://github.com/AI-S2-Lab/UniTalker 没内容

数据集

1.VOCASET VOCA

2.BIWI dataset Biwi 3D Audiovisual Corpus of Affective Communication

3.Flow-guided One-shot Talking Face Generation with a High-resolution Audio-visual Dataset  2021CVPR 网易伏羲

4.MMhead MM2025 https://openreview.net/pdf?id=L99kOQk12i

专门唱歌

1.SingAvatar: High-fidelity Audio-driven Singing Avatar Synthesis

ICME2024

文章https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10687925

说是会开源，实际没开源

2.MusicFace: Music-driven Expressive Singing Face Synthesis 上面有 没开源

数据集

1.SingingHead: A Large-scale 4D Dataset for Singing Head Animation

20240714 上海交大 https://openreview.net/profile?id=~Sijing_Wu1

文章https://arxiv.org/pdf/2312.04369

地址GitHub - wsj-sjtu/SingingHead: Official implentation of SingingHead: A Large-scale 4D Dataset for Singing Head Animation. (TMM 25)