This paper presents a method for many-to-one voice conversion using phonetic posteriorgrams (PPGs) based on an adversarial training of deep neural networks (DNNs). A conventional method for many-to-one VC can learn a mapping function from input acoustic features to target acoustic features through separately trained DNN-based speech recognition and synthesis models. However, 1) the differences among speakers observed in PPGs and 2) an over-smoothing effect of generated acoustic features degrade the converted speech quality. Our method performs a domain-adversarial training of the recognition model for reducing the PPG differences. In addition, it incorporates a generative adversarial network into the training of the synthesis model for alleviating the over-smoothing effect. Unlike the conventional method, ours jointly trains the recognition and synthesis models so that they are optimized for many-to-one VC. Experimental evaluation demonstrates that the proposed method significantly improves the converted speech quality compared with conventional VC methods.

This paper proposes Deep Neural Network (DNN)-based Voice Conversion (VC) using input-to-output highway networks. VC is a speech synthesis technique that converts input features into output speech parameters, and DNN-based acoustic models for VC are used to estimate the output speech parameters from the input speech parameters. Given that the input and output are often in the same domain (e.g., cepstrum) in VC, this paper proposes a VC using highway networks connected from the input to output. The acoustic models predict the weighted spectral differentials between the input and output spectral parameters. The architecture not only alleviates over-smoothing effects that degrade speech quality, but also effectively represents the characteristics of spectral parameters. The experimental results demonstrate that the proposed architecture outperforms Feed-Forward neural networks in terms of the speech quality and speaker individuality of the converted speech.

This paper describes a novel parameter generation algorithm for an HMM-based speech synthesis technique. The conventional algorithm generates a parameter trajectory of static features that maximizes the likelihood of a given HMM for the parameter sequence consisting of the static and dynamic features under an explicit constraint between those two features. The generated trajectory is often excessively smoothed due to the statistical processing. Using the over-smoothed speech parameters usually causes muffled sounds. In order to alleviate the over-smoothing effect, we propose a generation algorithm considering not only the HMM likelihood maximized in the conventional algorithm but also a likelihood for a global variance (GV) of the generated trajectory. The latter likelihood works as a penalty for the over-smoothing, i.e., a reduction of the GV of the generated trajectory. The result of a perceptual evaluation demonstrates that the proposed algorithm causes considerably large improvements in the naturalness of synthetic speech.