List your company

ResearchFormerly WizWand

Voice Conversion

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

CMU ARCTIC (test)

2.97Speaker Similarity

39

Jun 11, 2026

LibriSpeech Case 3: Source and Target > 1 min (test)

10WER

33

Jun 16, 2026

LibriSpeech Clean 100 Case 5: Source < 1 min, Target > 1 min

29Word Error Rate

33

Jun 16, 2026

0WER

27

Jun 4, 2026

LibriSpeech Clean 100 source and target longer than 1 min

0.78FAD

20

Jun 16, 2026

LibriSpeech English (test)

Palindromic Voice Conversion

0.717Speaker Similarity

20

Jun 9, 2026

LibriSpeech Case 2: Source and Target < 1 min (test)

0.23WER

19

Jun 16, 2026

LibriSpeech Case 4: source duration > 1 min, target duration < 1 min

41Word Error Rate

19

Jun 16, 2026

LibriSpeech (test-clean)

2.87WER

13

Feb 26, 2026

ELHE HE portion (160 unmodified utterances)

2.9CER

11

Feb 26, 2026

LibriTTS (test-clean)

2.04WER

11

Feb 26, 2026

Seed-TTS zh (test)

1.33WER

9

Apr 15, 2026

4.26nMOS

9

Feb 26, 2026

LibriTTS test-clean test-other and VCTK (merged)

3.668WER

8

Jun 23, 2026

LibriSpeech (test-clean and test-other)

2.18WER

8

Mar 11, 2026

SeedTTS VC Chinese (test)

1.25WER

8

Mar 6, 2026

SeedTTS VC English (test)

2.15WER

8

Mar 6, 2026

WavLM (continuous features)

0.721PPG Distance

7

Jun 8, 2026

7.26MCD

7

May 8, 2026

Seed-TTS en (test)

1.96WER

7

Apr 15, 2026

LibriSpeech (test-clean source, test-other target)

3.8MOS

7

Mar 11, 2026

Seed-TTS English

0.913SECS

7

Feb 26, 2026

Elliot Miller target speaker

3.22WER

7

Feb 26, 2026

LJSpeech target speaker

3.22WER

7

Feb 26, 2026

LibriSpeech short (test-clean)

90SIM

6

Jun 23, 2026