Direct Text to Speech Translation System using Acoustic Units

TL;DR

This paper introduces a novel direct text-to-speech translation system that bypasses intermediate transcriptions by using discrete acoustic units, leveraging a speech encoder, clustering, and vocoder, and demonstrates competitive results on a new multilingual corpus.

Contribution

It presents a new direct TTS translation framework using acoustic units, trained with multilingual models, achieving improved performance without relying on transcriptions.

Findings

Competitive performance on the CVSS corpus across language pairs

Significant improvement with multilingual pre-trained models

Effective use of acoustic units for direct translation

Abstract

This paper proposes a direct text to speech translation system using discrete acoustic units. This framework employs text in different source languages as input to generate speech in the target language without the need for text transcriptions in this language. Motivated by the success of acoustic units in previous works for direct speech to speech translation systems, we use the same pipeline to extract the acoustic units using a speech encoder combined with a clustering algorithm. Once units are obtained, an encoder-decoder architecture is trained to predict them. Then a vocoder generates speech from units. Our approach for direct text to speech translation was tested on the new CVSS corpus with two different text mBART models employed as initialisation. The systems presented report competitive performance for most of the language pairs evaluated. Besides, results show a remarkable improvement when initialising our proposed architecture with a model pre-trained with more languages.