Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Talking head synthesis research benchmarks · SOTA2 Research

Talking head synthesis

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

4.46Lip Sync Quality

18

Feb 26, 2026

May avatar Lieu audio

7.292Sync-D

10

May 29, 2026

May avatar Shaheen audio

12.012Sync-D

10

May 29, 2026

Conver-3D YouTube (test)

2.76FDD

9

Feb 26, 2026

Portrait Video Self-reconstruction (test)

35.15PSNR

8

May 29, 2026

Curated 5-Identity Audio-Visual Dataset (Macron, Paul, Obama, May, Stabenow) (test)

35.186PSNR

8

Mar 10, 2026

2.4LMD

6

May 12, 2026

21.586PSNR

6

Mar 13, 2026

Self-Collected Dataset 50 identities

32.77FID

6

Feb 26, 2026

VFHQ (first 100 frames)

26.22FID

6

Feb 26, 2026

3.994IQA

6

Feb 26, 2026

CelebV-HQ (test)

3.588IQA

6

Feb 26, 2026

HDTF Set B (test)

-3SyncNet (Offset)

6

Feb 26, 2026

HDTF Set A (test)

-4SyncNet Offset

6

Feb 26, 2026

Talking Head Synthesis Datasets (test)

Narrating For You

35.94PSNR

5

Feb 26, 2026

24.17PSNR

5

Feb 26, 2026

27PSNR

5

Feb 26, 2026

Popular datasets

HDTF May avatar User Study VFHQ CelebV-HQ

Follow for updates

Join SOTA2 Contact