SALTTS: Leveraging Self-Supervised Speech Representations for improved Text-to-Speech Synthesis
This is an incremental improvement for text-to-speech synthesis systems, addressing the need for richer speech representations.
The paper tackled the problem of limited speech representation in FastSpeech2 by using self-supervised learning (SSL) features to enhance text-to-speech synthesis, resulting in improved objective and subjective evaluation measures compared to the baseline.
While FastSpeech2 aims to integrate aspects of speech such as pitch, energy, and duration as conditional inputs, it still leaves scope for richer representations. As a part of this work, we leverage representations from various Self-Supervised Learning (SSL) models to enhance the quality of the synthesized speech. In particular, we pass the FastSpeech2 encoder's length-regulated outputs through a series of encoder layers with the objective of reconstructing the SSL representations. In the SALTTS-parallel implementation, the representations from this second encoder are used for an auxiliary reconstruction loss with the SSL features. The SALTTS-cascade implementation, however, passes these representations through the decoder in addition to having the reconstruction loss. The richness of speech characteristics from the SSL features reflects in the output speech quality, with the objective and subjective evaluation measures of the proposed approach outperforming the baseline FastSpeech2.