Towards zero-shot Text-based voice editing using acoustic context conditioning, utterance embeddings, and reference encoders
This work addresses the need for efficient and privacy-preserving voice editing tools for users in speech synthesis applications, though it is incremental as it builds on existing neural models.
The paper tackled the problem of text-based voice editing requiring costly and privacy-sensitive fine-tuning on target speaker data by proposing a zero-shot approach using pretrained speaker verification embeddings and a jointly trained reference encoder. The result was improved continuity of speaker identity and prosody in edited speech, as validated by subjective listening tests.
Text-based voice editing (TBVE) uses synthetic output from text-to-speech (TTS) systems to replace words in an original recording. Recent work has used neural models to produce edited speech that is similar to the original speech in terms of clarity, speaker identity, and prosody. However, one limitation of prior work is the usage of finetuning to optimise performance: this requires further model training on data from the target speaker, which is a costly process that may incorporate potentially sensitive data into server-side models. In contrast, this work focuses on the zero-shot approach which avoids finetuning altogether, and instead uses pretrained speaker verification embeddings together with a jointly trained reference encoder to encode utterance-level information that helps capture aspects such as speaker identity and prosody. Subjective listening tests find that both utterance embeddings and a reference encoder improve the continuity of speaker identity and prosody between the edited synthetic speech and unedited original recording in the zero-shot setting.