Prosodic Prominence and Boundaries in Sequence-to-Sequence Speech Synthesis

Abstract

Recent advances in deep learning methods have elevated synthetic speech quality to human level, and the field is now moving towards addressing prosodic variation in synthetic speech. Despite successes in this effort, the state-of-the-art systems fall short of faithfully reproducing local prosodic events that give rise to, e.g., word-level emphasis and phrasal structure. This type of prosodic variation often reflects long-distance semantic relationships that are not accessible for end-to-end systems with a single sentence as their synthesis domain. One of the possible solutions might be conditioning the synthesized speech by explicit prosodic labels, potentially generated using longer portions of text.In this work we evaluate whether augmenting the textual input with such prosodic labels capturing word-level prominence and phrasal boundary strength can result in more accurate realization of sentence prosody. We use an automatic wavelet-based technique to extract such labels from speech material, and use them as an input to a tacotron-like synthesis system alongside textual information.The results of objective evaluation of synthesized speech show that using the prosodic labels significantly improves the output in terms of faithfulness of f0 and energy contours, in comparison with state-of-the-art implementations.

DPP-TTS: Diversifying prosodic features of speech via determinantal point processes

“…Ye et al (2023) show that the prosody pattern reflecting relevant text context contributes to the enhanced TTS. Suni et al (2020); Nguyen et al (2020) suggest that prosodic boundary plays an important role in the naturalness and intelligibility of speech. To incorporate an inherent prosodic structure of input utterance within its context, we build Prosodic Boundary Detector (PBD) trained on a large corpus along with prominence labels.…”

ITAcotron 2: The Power of Transfer Learning in Expressive TTS Synthesis

“…Another controllable aspect is the speaker voice, introduced through additional speakeremebeddings extracted through a speaker verification network [13]. Finally, [27] proposed a methodology to control the prominence and boundaries by automatically deriving prosodic tags to augment the input character sequence. It is also possible to combine multiple techniques into a single conditioned architecture, as shown by [26].…”

A Modular Data-Driven Architecture for Empathetic Conversational Agents

“…This is the solution we decided to follow for the VC. Additionally, to further refine speech synthesis, we considered augmenting the text to be pronounced with prosodic clues (predicted by the empathetic controller) [20] useful, for example, for putting the focus on specific words.…”