Zero-Shot Multi-Speaker Text-To-Speech with State-Of-The-Art Neural Speaker Embeddings

Citations Over TimeTop 1% of 2020 papers

Abstract

While speaker adaptation for end-to-end speech synthesis using speaker embeddings can produce good speaker similarity for speakers seen during training, there remains a gap for zero-shot adaptation to unseen speakers. We investigate multi-speaker modeling for end-to-end text-to-speech synthesis and study the effects of different types of state-of-the-art neural speaker embeddings on speaker similarity for unseen speakers. Learnable dictionary encoding-based speaker embeddings with angular softmax loss can improve equal error rates over x-vectors in a speaker verification task; these embeddings also improve speaker similarity and naturalness for unseen speakers when used for zero-shot adaptation to new speakers in end-to-end speech synthesis.

Related Papers

• → A new Chinese text-to-speech system with high naturalness(2002)15 cited
• → Naturalness analysis of the speech synthesized by a TTS card(2016)2 cited
• → Speaker Recognition and Diarization(2010)3 cited
• Removing Preglottalization from Unit-Selection Synthesis: Towards the Linguistic Naturalness of Synthetic Czech Speech(2012)
• → Prosody Correction Preserving Speaker Individuality for Chinese-Accented Japanese HMM-Based Text-to-Speech Synthesis(2019)