The Deterministic Plus Stochastic Model of the Residual Signal and Its Applications

Abstract

Speech generated by parametric synthesizers generally suffers from a typical buzziness, similar to what was encountered in old LPC-like vocoders. In order to alleviate this problem, a more suited modeling of the excitation should be adopted. For this, we hereby propose an adaptation of the Deterministic plus Stochastic Model (DSM) for the residual. In this model, the excitation is divided into two distinct spectral bands delimited by the maximum voiced frequency. The deterministic part concerns the low-frequency contents and consists of a decomposition of pitch-synchronous residual frames on an orthonormal basis obtained by Principal Component Analysis. The stochastic component is a high-pass filtered noise whose time structure is modulated by an energy-envelope, similarly to what is done in the Harmonic plus Noise Model (HNM). The proposed residual model is integrated within a HMM-based speech synthesizer and is compared to the traditional excitation through a subjective test. Results show a significative improvement for both male and female voices. In addition the proposed model requires few computational load and memory, which is essential for its integration in commercial applications.

Excitation modeling for HMM-based speech synthesis: Breaking down the impact of periodic and aperiodic components

“…Experiments in [26] have shown that a mean glottal flow pulse (similar to eigenresidual in [17]) was rated better in quality than excitation using selection of natural pulses and equal to a pulse reconstructed from 12 PCA components. The latter comparison was also informally done using residual waveform in [17] with the same conclusion that using more components for modeling does not improve quality. In creaky voice synthesis [27], the type of deterministic waveform has also been shown to have relevant perceptual effect.…”

“…In [15], a hybrid approach makes use of a codebook of pitch-synchronous residual frames which are selected at synthesis time according to the down-sampled version of the excitation. In [16,17], the deterministic plus stochastic model (DSM) of the residual signal is proposed. DSM excitation consists of two components: the deterministic waveform called eigenresidual, which is obtained by principal component analysis (PCA) on a set of pitch-synchronous residual frames, and an aperiodic excitation delimited by maximum voiced frequency Fm and modulated in time according to a speaker-specific time envelope.…”

HMM-based speech synthesis with various degrees of articulation: A perceptual study

“…For the filter, we extracted the traditional Mel Generalized Cepstral (MGC) coefficients (with α = 0.42, γ = 0 and order of MGC analysis = 24). For the excitation, we used the Deterministic plus Stochastic Model (DSM) of the residual signal proposed in [11], since it was shown to significantly improve the naturalness of the delivered speech. More precisely, both deterministic and stochastic components of DSM were estimated on the training dataset for each degree of articulation.…”

Context-dependent acoustic modeling based on hidden maximum entropy model for statistical parametric speech synthesis

“…However, these advantages are achieved at the expense of one major disadvantage, i.e., degradation in the quality of synthetic speech [1]. This shortcoming results from three important factors: vocoding distortion [10][11][12][13], accuracy of statistical models [14][15][16][17][18][19][20][21][22][23][24][25], and accuracy of parameter generation algorithms [26][27][28]. This paper is an attempt to alleviate the second factor and improve the accuracy of statistical models.…”

“…In the training phase, first acoustic and contextual factors are extracted for the whole training database using a vocoder [12,29,30] and a natural language pre-processor. Next, the relationship between acoustic and contextual factors is modeled using a context-dependent statistical approach [14][15][16][17][18][19][20][21][22][23][24][25].…”