Voice Conversion from Non-parallel Corpora Using Variational Auto-encoder
This addresses a practical limitation in voice conversion applications by reducing dependency on scarce parallel data, though it is an incremental improvement over existing methods.
The paper tackles the problem of spectral conversion requiring parallel or aligned corpora by proposing a variational auto-encoder framework that enables training with non-parallel data, achieving competitive performance in objective and subjective evaluations compared to methods using aligned corpora.
We propose a flexible framework for spectral conversion (SC) that facilitates training with unaligned corpora. Many SC frameworks require parallel corpora, phonetic alignments, or explicit frame-wise correspondence for learning conversion functions or for synthesizing a target spectrum with the aid of alignments. However, these requirements gravely limit the scope of practical applications of SC due to scarcity or even unavailability of parallel corpora. We propose an SC framework based on variational auto-encoder which enables us to exploit non-parallel corpora. The framework comprises an encoder that learns speaker-independent phonetic representations and a decoder that learns to reconstruct the designated speaker. It removes the requirement of parallel corpora or phonetic alignments to train a spectral conversion system. We report objective and subjective evaluations to validate our proposed method and compare it to SC methods that have access to aligned corpora.