Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Audio-Based Linguistic Feature Extraction for Enhancing Multi-lingual and Low-Resource Text-to-Speech

Youngjae Kim, Yejin Jeon, Gary Geunbae Lee

TL;DR

Subjective and objective evaluations affirm the effectiveness of the approach for multi-lingual text-to-speech, and highlight its superiority in low-resource transfer learning for previously unseen language.

Abstract

The difficulty of acquiring abundant, high-quality data, especially in multi-lingual contexts, has sparked interest in addressing low-resource scenarios. Moreover, current literature rely on fixed expressions from language IDs, which results in the inadequate learning of language representations, and the failure to generate speech in unseen languages. To address these challenges, we propose a novel method that directly extracts linguistic features from audio input while effectively filtering out miscellaneous acoustic information including speaker-specific attributes like timbre. Subjective and objective evaluations affirm the effectiveness of our approach for multi-lingual text-to-speech, and highlight its superiority in low-resource transfer learning for previously unseen language.

Audio-Based Linguistic Feature Extraction for Enhancing Multi-lingual and Low-Resource Text-to-Speech

TL;DR

Subjective and objective evaluations affirm the effectiveness of the approach for multi-lingual text-to-speech, and highlight its superiority in low-resource transfer learning for previously unseen language.

Abstract

The difficulty of acquiring abundant, high-quality data, especially in multi-lingual contexts, has sparked interest in addressing low-resource scenarios. Moreover, current literature rely on fixed expressions from language IDs, which results in the inadequate learning of language representations, and the failure to generate speech in unseen languages. To address these challenges, we propose a novel method that directly extracts linguistic features from audio input while effectively filtering out miscellaneous acoustic information including speaker-specific attributes like timbre. Subjective and objective evaluations affirm the effectiveness of our approach for multi-lingual text-to-speech, and highlight its superiority in low-resource transfer learning for previously unseen language.
Paper Structure (17 sections, 3 equations, 2 figures, 3 tables)

This paper contains 17 sections, 3 equations, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Preliminaries
  4. Audio-Based Language Encoder
  5. Training Protocol
  6. Experiments
  7. Datasets
  8. Multi-lingual Experimental Settings
  9. Low-resource Experimental Settings
  10. Results and Discussion
  11. Multi-lingual Results
  12. Low-resource Result
  13. Ablation Study
  14. Conclusion
  15. Limitations
  16. ...and 2 more sections

Figures (2)

  • Figure 1: Conceptual visualization comparing prior multi-lingual methodologies that utilize language IDs for language representations (upper) with the proposed methodology (lower). The architecture after the text encoder follows VITS, which we omit for brevity.
  • Figure 2: Distribution of language embeddings with and without the projection layer. Visualizations were conducted using 2D PCA. Utilization of an addition projection layer results in distinct language-specific clusters.