A statistical method for estimation of speech spectral density and formant frequencies

By F. Itakura. 1970

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Speech & Audio

Citations in the corpus

WaveNet: A Generative Model for Raw Audio (2016, methodology): “Extracting vocoder parameters can be viewed as estimation of a generative model parameters given speech signals (Itakura & Saito, 1970; Imai & Furuichi, 1988).”
Mel-generalized cepstral analysis - a unified approach to speech spectral estimation (1994, methodology): “Linear prediction [1], [2] is a generally accepted method for obtaining all-pole representation of speech.”
Statistical Parametric Speech Synthesis (2007, methodology): “Estimate model parameters based on ML c = arg max c p(x | c) • p(x | c): AR model → Linear predictive analysis [5] • p(x | c): EX model → (ML-based) cepstral analysis [6]”
Restructuring speech representations using a pitch-adaptive time-frequency smoothing and an instantaneous-frequency-based F0 extraction: Possible role of a repetitive structure in sounds (1999, background): “The channel VOCODER and its modern variant, LPC (Linear Predictive Coding) [14] [4] are potentially very flexible in parameter manipulations , because there are little inter-related constraints between spectral and source parameters.”
Automatic Speech Recognition - A Brief History of the Technology Development (2004, background): “By the mid 1970's, the basic ideas of applying fundamental pattern recognition technology to speech recognition, based on LPC methods, were proposed by Itakura [22], Rabiner and Levinson [23] and others.”
Introduction to Digital Speech Processing (2007, methodology): “Itakura and Saito [53, 54], showed that the parameters ki in the Levinson–Durbin recursion and the lattice filter interpretation obtained from it also could be derived by looking at linear predictive analysis from a statistical perspective.”
Directly modeling speech waveforms by neural networks for statistical parametric speech synthesis (2015, unknown): “Extracting a parametric representation of speech can also be viewed as ML estimation of the model parameters given the wave form [9] [10].”
Integration of Spectral Feature Extraction and Modeling for HMM-Based Speech Synthesis (2014, methodology): “It should be noted that the spectral evaluation function of mel-cepstral analysis has the same form as that of LPC analysis [26].”
Excitation source analysis for high-quality speech manipulation systems based on an interference-free representation of group delay with minimum phase response compensation (2014, methodology): “Source ﬁlter models provide a basis for VOCODER [12], LPC [13, 14], Cepstrum-based models [15, 16, 17] and STRAIGHT 1 [1, 2].”
Statistical parametric speech synthesis: from HMM to LSTM-RNN (2015, methodology): “Estimate model parameters based on ML c = arg max c p(x | c) • p(x | c): AR model → Linear predictive analysis [5] • p(x | c): EX model → (ML-based) cepstral analysis [6]”
Statistical approach to vocal tract transfer function estimation based on factor analyzed trajectory HMM (2008, background): “Itakura and Saito [1] proposed maximum likelihood estimation of a speech spectral envelope.”