Hierarchical Prosody Modeling for Non-Autoregressive Speech Synthesis
This work addresses prosody modeling for speech synthesis, offering a method to enhance control and naturalness in synthesized speech, though it appears incremental as it builds on existing non-autoregressive TTS frameworks.
The paper tackled the challenge of predicting natural prosody in non-autoregressive text-to-speech synthesis by proposing a hierarchical architecture that conditions phoneme-level prosody on word-level features, resulting in improved audio quality and prosody naturalness in evaluations.
Prosody modeling is an essential component in modern text-to-speech (TTS) frameworks. By explicitly providing prosody features to the TTS model, the style of synthesized utterances can thus be controlled. However, predicting natural and reasonable prosody at inference time is challenging. In this work, we analyzed the behavior of non-autoregressive TTS models under different prosody-modeling settings and proposed a hierarchical architecture, in which the prediction of phoneme-level prosody features are conditioned on the word-level prosody features. The proposed method outperforms other competitors in terms of audio quality and prosody naturalness in our objective and subjective evaluation.