In this paper we propose a deep neural network to model the conditional probability of the spectral differences between natural and synthetic speech. This allows us to reconstruct the spectral fine structures in speech generated by HMMs. We compared the new stochastic data-driven postfilter with global variance based parameter generation and modulation spectrum enhancement. Our results confirm that the proposed method significantly improves the segmental quality of synthetic speech compared to the conventional methods.
Bibliographic reference. Chen, Ling-Hui / Raitio, Tuomo / Valentini-Botinhao, Cassia / Yamagishi, Junichi / Ling, Zhen-Hua (2014): "DNN-based stochastic postfilter for HMM-based speech synthesis", In INTERSPEECH-2014, 1954-1958.