Closed phase covariance analysis based on constrained linear prediction for glottal inverse filtering

Abstract

Closed phase (CP) covariance analysis is a widely used glottal inverse filtering method based on the estimation of the vocal tract during the glottal CP. Since the length of the CP is typically short, the vocal tract computation with linear prediction (LP) is vulnerable to the covariance frame position. The present study proposes modification of the CP algorithm based on two issues. First, and most importantly, the computation of the vocal tract model is changed from the one used in the conventional LP into a form where a constraint is imposed on the dc gain of the inverse filter in the filter optimization. With this constraint, LP analysis is more prone to give vocal tract models that are justified by the source-filter theory; that is, they show complex conjugate roots in the formant regions rather than unrealistic resonances at low frequencies. Second, the new CP method utilizes a minimum phase inverse filter. The method was evaluated using synthetic vowels produced by physical modeling and natural speech. The results show that the algorithm improves the performance of the CP-type inverse filtering and its robustness with respect to the covariance frame position.

Modeling the effects of a posterior glottal opening on vocal fold dynamics with implications for vocal hyperfunction

“…It is noted that the addition of the gap yields smooth contours of the glottal airflow during the closed portion of the cycle, which is in agreement with experimental observations from inverse filtering in human subjects. 8,32 However, it is noted that this smoothing of the glottal airflow may occur from multiple other conditions and is not uniquely tied to the PGO effects.…”

“…1. This model is an extension of the classical two-mass model 28 that better represents physiological aspects of the vocal folds, and has been used to study source-filter interaction, 29,30 voice pathologies, 19,31 inverse filtering, 32 and muscle activation, 33 among others. The equations of motion for the vocal fold model are taken exactly as in the original paper 27 and are thus omitted for brevity.…”

A Fast Method for High-Resolution Voiced/Unvoiced Detection and Glottal Closure/Opening Instant Estimation of Speech

“…This provides a smooth approximation of the GFD during voiced speech. Of course there are much more accurate GFD estimation methods in the literature [8]- [11]. However, our aim here is to obtain a fast and convenient approximation of the GFD in the context of GEFBA.…”

“…In speech analysis nomenclature, these timing instants are called glottal closure instants (GCIs) and glottal opening instants (GOIs). Applications of GCI and GOI estimation are numerous, including pitch tracking [1], [2], voice source modeling [3]- [6], speech enhancement [7], closedphase analysis and glottal flow estimation [8]- [11], speaker identification [9], [12], [13], speech dereverberation [14], speech synthesis [15], [16], speech coding [17], speech modification [18], [19] and speech transformations [20].…”

Development of a glottal area index that integrates glottal gap size and open quotient

“…One way of recovering the voice source signal from the acoustic sound pressure or oral airflow signal is via inverse filtering, which removes the vocal tract filtering effect. Inverse filtering is sensitive to recording conditions and experimental setup, and several methods have been proposed (e.g., Rothenberg, 1973;Javkin et al, 1987;Alku, 1992;Alku et al, 2006Alku et al, , 2009. A more direct way of observing vocal fold vibration is through high-speed video recording.…”