Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Lip-to-speech synthesis on LRS3 unseen speaker protocol (test)
Loading...
21.9
WER
LipVoicerofficial
21.876
22.038
22.2
22.362
Feb 2, 2026
WER
STOI
DNSMOS
LSE-D
LSE-C
Updated 4d ago
Evaluation Results
Method
Method
Links
WER
STOI
DNSMOS
LSE-D
LSE-C
LipVoicerofficial
variant=official, voco...
2026.02
21.9
93
3.19
8.166
6.308
LipVoicerrecon
variant=reimplementati...
2026.02
22.1
93
3.17
8.516
6.4
LipSody
prosody_modeling=explicit
2026.02
22.5
92
3.18
8.435
6.404
Feedback
Search any
task
Search any
task