Pororo TTS - Samples

Audio samples of the open-source library PORORO: Platform Of neuRal mOdels for natuRal language prOcessing. Pororo TTS features 11 multi-language support, Cross-lingual Voice Style Transfer, and Code-Switching functions.

Contents:

• Multilingual TTS
• Cross-lingual Voice Style Transfer
• Code Switching

Notes:

• These samples can be reproduced at https://github.com/kakaobrain/pororo.
• Synthesis text to speech using trained model. Output audio's sample rate is 22050.
• Samples shown here were selected based on diversity and quality.

Multilingual TTS

Samples generated by Pororo-TTS. Supports 11 language. (English, Chinese, Korean, Japanese, Jejueo, French, Dutch, German, Russian, Spanish, Finnish). It was trained using CSS10, KSS, and JSS datasets.

Samples

Korean Samples

English Samples

Chinese Samples

Japenses Samples

Jejueo Samples

French Samples

German Samples

Finnish Samples

Dutch Samples

Spanish Samples

Russian Samples

Cross-lingual Voice Style Transfer

Samples generated by Pororo-TTS. Cross-lingual voice style transfer function allows the speaker to speak other languages.

Samples

Korean speaker reads Japanese & Korean speaker reads Chinese

English speaker reads Korean & Japanese speaker reads Korean

German speaker reads French & German speaker reads Dutch

French speaker reads German & French speaker reads Korean

Dutch speaker reads English & Dutch speaker reads German

Code Switching

Samples generated by Pororo-TTS. Code-switching text-to-speech aims to enable a system to speak more than two speakers in the same utterance. ^* indicates the language of the text.

Samples

Speakers: English^*, German / Speakers: German, Korean^*

Speakers: German^*, Chinese / Speakers: Dutch^*, German

Speakers: Japanese^*, Dutch, English / Speakers: French^*, German