Refined WaveNet Vocoder for Variational Autoencoder Based Voice Conversion
This work addresses a specific bottleneck in voice conversion systems for speech processing applications, but it is incremental as it refines an existing method.
The paper tackled the quality distortion in voice conversion caused by mismatched training and testing data for WaveNet vocoders, and the proposed refinement framework using VAE self-reconstructed features improved performance as shown in objective and subjective experiments.
This paper presents a refinement framework of WaveNet vocoders for variational autoencoder (VAE) based voice conversion (VC), which reduces the quality distortion caused by the mismatch between the training data and testing data. Conventional WaveNet vocoders are trained with natural acoustic features but conditioned on the converted features in the conversion stage for VC, and such a mismatch often causes significant quality and similarity degradation. In this work, we take advantage of the particular structure of VAEs to refine WaveNet vocoders with the self-reconstructed features generated by VAE, which are of similar characteristics with the converted features while having the same temporal structure with the target natural features. We analyze these features and show that the self-reconstructed features are similar to the converted features. Objective and subjective experimental results demonstrate the effectiveness of our proposed framework.