Towards Developing State-of-the-Art TTS Synthesisers for 13 Indian Languages with Signal Processing aided Alignments

TL;DR

This paper presents a novel approach combining signal processing cues with forced alignments to improve duration modeling in end-to-end TTS systems, resulting in high-quality speech synthesis for 13 Indian languages, especially in low-resource settings.

Contribution

It introduces a signal processing-aided alignment method that enhances duration accuracy and synthesis quality in multilingual TTS systems, outperforming existing approaches.

Findings

Systems outperform other alignment methods in low-resource scenarios.

Proposed approach achieves better synthesis quality than existing systems.

Outperforms current best TTS systems for 13 Indian languages.

Abstract

End-to-end (E2E) systems synthesise high-quality speech, but this typically requires a large amount of data. As E2E synthesis progressed from Tacotron to FastSpeech2, it became evident that features representing prosody, particularly sub-word durations, are important for error-free synthesis. Variants of FastSpeech use a teacher model or forced alignments for training. This paper uses signal processing cues in tandem with forced alignment to produce accurate phone boundaries for the training data. As a result of better duration modelling, good-quality synthesisers are developed. Evaluations indicate that systems developed using the proposed signal processing-aided approach are better than systems developed using other alignment approaches, especially in low-resource scenarios. Our systems also outperform the existing best TTS systems available for 13 Indian languages.