List your company

System naturalness prediction on ConvTTS (test)

0.57PCCConversational Speech Naturalness Predictor (Whisper)
0.321440.385970.45050.51503Mar 2, 2026

Evaluation results

MethodLinks
Conversational Speech Naturalness Predictor (Whisper)Encoder=Whisper, Input Channel(s)=dual2026.030.570.560.218
Conversational Speech Naturalness Predictor (Whisper)Encoder=Whisper, Input Channel(s)=single2026.030.550.5380.226
Conversational Speech Naturalness Predictor (AES)Encoder=AES, Input Channel(s)=dual2026.030.5420.5340.227
Conversational Speech Naturalness Predictor (AES)Encoder=AES, Input Channel(s)=single2026.030.5230.5230.234
Conversational Speech Naturalness Predictor (WavLM)Encoder=WavLM, Input Channel(s)=dual2026.030.5170.5220.236
Conversational Speech Naturalness Predictor (WavLM)Encoder=WavLM, Input Channel(s)=single2026.030.4880.4910.246
Conversational Speech Naturalness Predictor (Whisper)Encoder=Whisper, Input Channel(s)=dual2026.030.4820.4510.208
Conversational Speech Naturalness Predictor (AES)Encoder=AES, Input Channel(s)=dual2026.030.4550.430.213
Conversational Speech Naturalness Predictor (Whisper)Encoder=Whisper, Input Channel(s)=single2026.030.4330.4060.221
Conversational Speech Naturalness Predictor (WavLM)Encoder=WavLM, Input Channel(s)=dual2026.030.4090.3580.227
Conversational Speech Naturalness Predictor (AES)Encoder=AES, Input Channel(s)=single2026.030.3840.3520.228
Conversational Speech Naturalness Predictor (WavLM)Encoder=WavLM, Input Channel(s)=single2026.030.3310.2790.242