Alternate Endings: Improving Prosody for Incremental Neural TTS with Predicted Future Text Input
Preprint
Abstract
The prosody of a spoken word is determined by its surrounding context. In incremental text-to-speech synthesis, where the synthesizer produces an output before it has access to the complete input, the full context is often unknown which can result in a loss of naturalness in the synthesized speech. In this paper, we investigate whether the use of predicted future text can attenuate this loss. We compare several test conditions of next future word: (a) unknown (zero-word), (b) language model predicted, (c) randomly predicted and (d) ground-truth. We measure the prosodic features (pitch, energy and duration) and find that predicted text provides significant improvements over a zero-word lookahead, but only slight gains over random-word lookahead. We confirm these results with a perceptive test.Keywords
All Related Versions
This publication has 9 references indexed in Scilit:
- Incremental Text-to-Speech Synthesis Using Pseudo Lookahead With Large Pretrained Language ModelIEEE Signal Processing Letters, 2021
- Incremental Machine Speech Chain Towards Enabling Listening While Speaking in Real-Time2020
- Espnet-TTS: Unified, Reproducible, and Integratable Open Source End-to-End Text-to-Speech Toolkit2020
- Parallel Wavegan: A Fast Waveform Generation Model Based on Generative Adversarial Networks with Multi-Resolution Spectrogram2020
- Neural iTTS: Toward Synthesizing Speech in Real-time with End-to-end Neural Text-to-Speech Framework2019
- Introducing Parselmouth: A Python interface to PraatJournal of Phonetics, 2018
- Limits on lexical prediction during readingCognitive Psychology, 2016
- On the just noticeable difference for tempo in speechJournal of Phonetics, 2007
- Differential sensitivity to pitch distance, particularly in speechThe Journal of the Acoustical Society of America, 1981