Incremental Text To Speech For Neural Sequence-to-sequence Models Using Reinforcement Learning

Abstract

Modern approaches to text to speech require the entire input character sequence to be processed before any audio is synthesised. This latency limits the suitability of such models for time-sensitive tasks like simultaneous interpretation. Interleaving the action of reading a character with that of synthesising audio reduces this latency. However, the order of this sequence of interleaved actions varies across sentences, which raises the question of how the actions should be chosen. We propose a reinforcement learning based framework to train an agent to make this decision. We compare our performance against that of deterministic, rule-based systems. Our results demonstrate that our agent successfully balances the trade-off between the latency of audio generation and the quality of synthesised audio. More broadly, we show that neural sequence-to-sequence models can be adapted to run in an incremental manner.

Stats

• citations4
• S2 citations—
• github stars0
• HF likes0
• heat score5.24
• arxiv keymohan2020incremental

Related papers

• Incremental Text-to-speech Synthesis Using Pseudo Lookahead With Large Pretrained Language Model (2020)8.82
• Incremental Text-to-speech Synthesis With Prefix-to-prefix Framework (2019)7.81
• Sequence-to-sequence ASR Optimization Via Reinforcement Learning (2017)9.41
• Sequence-to-sequence Learning Via Attention Transfer For Incremental Speech Recognition (2020)4.52
• What The Future Brings: Investigating The Impact Of Lookahead For Incremental Neural TTS (2020)6.34
• Low-latency Incremental Text-to-speech Synthesis With Distilled Context Prediction Network (2021)6.34
• Reinforce-aligner: Reinforcement Alignment Search For Robust End-to-end Text-to-speech (2021)8.09
• Iterative Autoregression: A Novel Trick To Improve Your Low-latency Speech Enhancement Model (2022)5.24