Automatic Modeling of Acoustic Perception of Breathiness in Pathological Voices

Abstract

This paper revisits the modeling of acoustic perceptions of breathy voice (BV) quality for automatic assessment of perturbations in pathologic speech. Several acoustic measures related with the signal periodicity, harmonic components, and aspiration noise are studied to predict breathiness judgments performed on sustained vowel phonations. A novel comprehensive automatic measure is proposed that provides the highest correlation index (88.5%) with breathiness judgment performed by trained specialists on simulated and recorded utterances. The new measure reveals the most relevant aspects of BV quality and provides a vehicle to obtain reliable objectives judgments of such speech perturbation.

Relation of Structural and Vibratory Kinematics of the Vocal Folds to Two Acoustic Measures of Breathy Voice Based on Computational Modeling

“…These findings are in contrast to those of Castillo-Guerra and Ruíz (2009), who compared acoustic measures to perceptual judgments of breathy voice. Two measures were used to assess the ratio between H1 and H2: harmonic energy (HE) which was based on the radiated acoustic signal, and harmonic energy of residue (HE RES ), which was calculated after passage through an inverse filter model.…”

“…Two measures were used to assess the ratio between H1 and H2: harmonic energy (HE) which was based on the radiated acoustic signal, and harmonic energy of residue (HE RES ), which was calculated after passage through an inverse filter model. The measure HE RES better accounted for variance in the perceptual judgments across multiple types of perturbation, leading the authors to conclude that filtering effects of the vocal tract from the signal improves the accuracy of the information conveyed about breathy voice (Castillo-Guerra & Ruíz, 2009). The most likely explanation for this is that they did not appear to correct the harmonic amplitudes for the influences of F1 in their calculation.…”

Speech emotion recognition using emotion perception spectral feature

“…Breath feature is a vocal feature according to the mechanism of speech production, 8 which includes periodic perturbation coefficient, harmonic‐noise ratio, glottic‐noise excitation ratio, harmonic structure energy, harmonic energy residual, and harmonic‐signal ration.…”

“…Tone quality feature generally compose of formant frequency, frequency perturbation, and amplitude perturbation 7 . Castillo‐Guerra and Ruiz 8 studied the automatic evaluation model to the quality of acoustic breathing and proposed a tone quality feature for recognizing speech emotion, the feature consists of periodic disturbance coefficient, amplitude disturbance coefficient, harmonic noise ratio, glottic noise excitation ratio, harmonic energy, harmonic energy residue, and harmonic signal ratio. Tone quality feature has a high recognition performance in vigorous active emotions, such as anger and surprise.…”

An efficient approach using HOS-based parameters in the LPC residual domain to classify breathy and rough voices

“…In this paper, traditional acoustic measures concerning the signal periodicity, harmonic components, and aspiration noise were calculated in addition to harmonic energy of residue (HE RES ), harmonic-to-signal ratio (HSR), and number of voiced frames (NVF). The best classification performance (88.5%) was achieved with a best subset regression (BSR) analysis using linear combinations of multiple parameters [15]. Wester suggested two methods to automatically classify voice quality: regression analysis and hidden markov models (HMMs).…”