Chatterbox Multilingual rows in a 5-language voice-cloning benchmark
#61
by aufklarer - opened
I included Chatterbox Multilingual fp16 in a local voice-cloning benchmark across English, German, Modern Standard Arabic, and Spanish:
https://www.soniqo.audio/blog/voice-cloning-benchmarks
Chatterbox was competitive on Arabic and Spanish in this run. I did not include a Chinese row yet because the current Swift tokenizer path does not support that benchmark path cleanly.
The benchmark uses Google FLEURS references and includes reference audio, generated audio, speaker similarity, WER/CER, generated audio length, and RTF for each row.
This is an engineering benchmark rather than a MOS study, but I wanted to share the Chatterbox rows with the model community.