Automatic Voice Identification after Speech Resynthesis using PPG
This addresses the challenge of speaker identity preservation in voice conversion and speech editing for applications like media monitoring, but it is incremental as it builds on existing PPG methods.
The paper tackles the problem of disentangling speaker and phonetic content in speech resynthesis using Phonetic PosteriorGrams (PPG), and demonstrates that an automatic speaker verification model cannot recover the source speaker after resynthesis, even when trained on synthetic data.
Speech resynthesis is a generic task for which we want to synthesize audio with another audio as input, which finds applications for media monitors and journalists.Among different tasks addressed by speech resynthesis, voice conversion preserves the linguistic information while modifying the identity of the speaker, and speech edition preserves the identity of the speaker but some words are modified.In both cases, we need to disentangle speaker and phonetic contents in intermediate representations.Phonetic PosteriorGrams (PPG) are a frame-level probabilistic representation of phonemes, and are usually considered speaker-independent.This paper presents a PPG-based speech resynthesis system.A perceptive evaluation assesses that it produces correct audio quality.Then, we demonstrate that an automatic speaker verification model is not able to recover the source speaker after re-synthesis with PPG, even when the model is trained on synthetic data.