Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

LibriTTS

Benchmarks

Task NameDataset NameSOTA ResultTrend
Speech ReconstructionLibriTTS clean (test)
PESQ4.186
50
Speech ReconstructionLibriTTS (test-other)
UTMOS3.91
44
Audio GenerationLibriTTS (dev)
M-STFT1.3647
18
Speech SynthesisLibriTTS (test)
MOS4.9134
17
Text-to-SpeechLibriTTS (test)
MOS4.54
16
Text-to-SpeechLibriTTS clean (test)
WER0.018
15
Text-to-SpeechLibriTTS zero-shot
UTMOS4.3026
14
Waveform GenerationLibriTTS 24,000 Hz (test)
UTMOS3.7229
13
Waveform GenerationLibriTTS (dev)
M-STFT1.2129
12
Voice ConversionLibriTTS (test-clean)
WER2.04
11
Speech SynthesisLibriTTS 24,000 Hz (test)
MOS4.28
11
Waveform GenerationLibriTTS-R clean (test)
Speech BERT Score100
10
Audio ReconstructionLibriTTS clean (test)
Mel Distance0.3442
10
VocodingLibriTTS (dev-other)
MAE0.0986
10
Neural VocodingLibriTTS clean (dev)
MAE0.0931
10
Speech SynthesisLibriTTS (ID)
PESQ4.5
9
Audio WatermarkingLibriTTS
PESQ4.3289
8
Generative Speech WatermarkingLibriTTS OOD (test)
STOI0.9789
8
Neural VocodingLibriTTS
UTMOS4.058
8
Speech ResynthesisLibriTTS (test-clean)
WER3.32
7
Speech ReconstructionLibriTTS (test)
PESQ4.16
7
Universal Neural VocodingLibriTTS clean and other (dev)
M-STFT0.7997
6
Voice ConversionLibriTTS unseen-to-unseen (test-clean)
MOS4.27
6
Voice ConversionLibriTTS to VCTK (unseen-to-seen) (test-clean)
MOS4.29
6
Speech DereverberationLibriTTS (clean part) + openSLR26/28 RIR (test)
PESQ2.87
5
Showing 25 of 42 rows