Integration of Spectral Feature Extraction and Modeling for HMM-Based Speech Synthesis

Abstract

SUMMARYThis paper proposes a novel approach for integrating spectral feature extraction and acoustic modeling in hidden Markov model (HMM) based speech synthesis. The statistical modeling process of speech waveforms is typically divided into two component modules: the frame-byframe feature extraction module and the acoustic modeling module. In the feature extraction module, the statistical mel-cepstral analysis technique has been used and the objective function is the likelihood of mel-cepstral coefficients for given speech waveforms. In the acoustic modeling module, the objective function is the likelihood of model parameters for given melcepstral coefficients. It is important to improve the performance of each component module for achieving higher quality synthesized speech. However, the final objective of speech synthesis systems is to generate natural speech waveforms from given texts, and the improvement of each component module does not always lead to the improvement of the quality of synthesized speech. Therefore, ideally all objective functions should be optimized based on an integrated criterion which well represents subjective speech quality of human perception. In this paper, we propose an approach to model speech waveforms directly and optimize the final objective function. Experimental results show that the proposed method outperformed the conventional methods in objective and subjective measures. key words: integrative model, HMM-based speech synthesis, acoustic modeling, mel-cepstral analysis, trajectory HMM

Speech Synthesis Based on Deep Neural Networks with Direct Modeling of Amplitude Spectra

“…Prior to the recent use of neural networks in TTS, some authors tried to implement the end-to-end concept, sometimes regarded as waveform models [21], [22].…”

Directly modeling speech waveforms by neural networks for statistical parametric speech synthesis

“…(14). By back-propagating the derivative of the log likelihood function through the network, the network weights can be updated to maximize the log likelihood.…”

“…the log spectral distortion-version of minimum generation error training (MGE-LSD) [11], statistical vocoder (STAVOCO) [12], waveform-level statistical model [13], and mel-cepstral analysis-integrated hidden Markov models (HMMs) [14]. However, there are limitations in these approaches, such as the use of spectra rather than waveforms, the use of overlapping and shifting frames as unit, and fixing decision trees [15], which represent the mapping from linguistic features to acoustic ones [16].…”