Towards Expressive Speaking Style Modelling with Hierarchical Context Information for Mandarin Speech Synthesis
This addresses the issue of lacking speech variations in Mandarin speech synthesis for applications like lectures, but it is incremental as it builds on existing expressive synthesis methods.
The paper tackled the problem of inflexible speaking style in expressive speech synthesis by proposing a hierarchical framework to model style from context, including inter-phrase and inter-sentence relations, and demonstrated significant improvements in naturalness and expressiveness on a Mandarin lecture dataset.
Previous works on expressive speech synthesis mainly focus on current sentence. The context in adjacent sentences is neglected, resulting in inflexible speaking style for the same text, which lacks speech variations. In this paper, we propose a hierarchical framework to model speaking style from context. A hierarchical context encoder is proposed to explore a wider range of contextual information considering structural relationship in context, including inter-phrase and inter-sentence relations. Moreover, to encourage this encoder to learn style representation better, we introduce a novel training strategy with knowledge distillation, which provides the target for encoder training. Both objective and subjective evaluations on a Mandarin lecture dataset demonstrate that the proposed method can significantly improve the naturalness and expressiveness of the synthesized speech.