Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Voice Conversion on Expresso OOD
Loading...
0.543
F0 Correlation
Voc-only
0.37556
0.41903
0.4625
0.50597
Jan 27, 2026
F0 Correlation
Speaker Similarity (SpkSim)
UTMOS Score
Word Error Rate (WER)
Updated 4d ago
Evaluation Results
Method
Method
Links
F0 Correlation
Speaker Similarity (SpkSim)
UTMOS Score
Word Error Rate (WER)
Voc-only
alpha=1
2026.01
0.543
0.608
2.96
26.8
Phonological Tokenizer
alpha=0.1
2026.01
0.538
0.724
3.58
12.6
WavTokenizer
Type=acoustic
2026.01
0.52
0.352
2.24
27.7
ASR-only
alpha=0
2026.01
0.391
0.738
3.61
12.6
SpeechTokenizer
Type=hybrid
2026.01
0.388
0.706
3.13
24
Discrete WavLM
Type=phonetic
2026.01
0.382
0.737
3.47
12.2
Feedback
Search any
task
Search any
task