This paper presents a novel method of formant tracking. It shows that formant information can be extracted from the cepstrum coefficients . Explicit nonlinear formulas has been developed to map psd (power spectral density) of speech signal to formant frequencies. Formants can be tracked by ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques). An ”‘constrained”’ ESPRIT s...

This paper presents a novel approach to integration of formant frequency and conventional MFCC data in phone recognition experiments on TIMIT. Naive use of format data introduces classification errors if formant frequency estimates are poor, resulting in a net drop in performance. However, by exploiting a measure of confidence in the formant frequency estimates, formant data can contribute to c...

The speech intelligibility of Mandarin Chinese sentences of various spectral regions, regarding band-stop conditions (one or two “holes” in the spectrum), was investigated through subjective listening tests. Results demonstrated significant effects on Mandarin Chinese sentence intelligibilities when a single or a pair of spectral holes was introduced. Meanwhile, it revealed the importance of th...

Formant frequencies were determined from vocalic utterances of a prelinguistically deafened child implanted with a Nucleus 22-channel device at age 5 years. Speech samples were obtained from recordings of speech made 5 and 2 months prior to implantation and at 6, 12, 18, 24, and 36 months post implantation. Prior to implantation, first formant values showed a greater range than those of normal ...

This paper concerns Multiple-level Segmental HiddenMarkov Models (M-SHMMs) in which the relationship between symbolic and acoustic representations of speech is regulated by a formant-based intermediate representation. New TIMIT phone recognition results are presented, confirming that the theoretical upper-bound on performance is achieved provided that either the intermediate representation or t...

In this paper we present a new formant tracking algorithm where the formant frequencies estimation was based on local maxima detection of a time frequency representation. This representation can be shown by a scalogram issued from a complex wavelet transform. The formant frequency candidates are validated as local maxima of scalogram which correspond to wavelet ridges. Then in the proposed algo...

The goal of the work described in this paper is to develop a very low bit rate vocoding scheme. The vocoder is a typical LPC vocoder, whose parameters are post-processed on a phone-byphone basis, resulting in a variable bit rate segment vocoder. Given the well known speaker recognizability problems presented by vocoders at such low bit rates, we have attempted to integrate a speaker modificatio...

In recent years, several studies have shown the relationship between snoring and obstructive sleep apnea syndrome (OSAS). Instead of time domain analysis of snoring signal, the spectral features and shapes of snores have been found different in simple snorers and OSAS patients. In this study, we propose a method to differentiate simple snorers and OSAS patients based on spectral envelope estima...

The objective of this paper is to critically evaluate the performance of a nonstationary analysis method in tracking speech formant frequencies as they change with time due to the natural variations in the vocal-tract system during speech production. The method of instantaneous frequency estimation is applied to the tracking of speech formant frequencies to observe the time variations in the vo...