Short-Term Stability Measures for the Evaluation of Vocal Quality

Abstract

The vocal quality of 64 normal subjects and 57 subjects suffering various degrees of glottal cancer was investigated using acoustic measures of six different aspects of the voice signal: tone period perturbation, amplitude perturbation, waveform perturbation, vocal noise, spectral periodicity and spectral distortion. The measures were estimated taking the glottal cycle as temporal reference unit to make the influence of the differences in tone period from one person to another as low as possible. The measures were evaluated with regard to (a) their ability to discriminate between healthy and sick subjects, and (b) their correlation with the perceptual evaluation of four trained listeners. The results suggest that signal processing techniques are unsatisfactory for clinical diagnoses but useful for monitoring voice quality.

The perceptual structure of pathologic voice quality

“…For example, many believe that jitter has consistent perceptual correlates ͑roughness is the most frequently mentioned; Gerratt and Kreiman, 1995͒, and thus that it is a good measure of vocal quality. However, correlations between measures of jitter and perceived roughness have ranged from Ϫ0.01 ͑Es-kenazi et al, 1990͒ to 0.98 ͑Heiberger and Horii, 1982͒; and jitter has been significantly ͑if not always highly͒ associated with many qualities, including grade/severity of pathology ͑Askenfelt and Wolfe andRatusnik, 1988͒, breathiness ͑Hirano et al, 1988;Eskenazi et al, 1990;Feijoo andHernandez, 1990͒, hoarseness ͑Haji et al, 1986;Horiguchi et al, 1987͒, andstrain ͑Arends et al, 1990͒.…”

Computer Aided Recognition of Vocal Folds Disorders by Means of RASTA-PLP

“…The term mel refers to a kind of measurement related to perceived frequency. The mapping between the real frequency scale (Hz) and the perceived frequency scales (mels) is approximately linear below 1 kHz and logarithmic at higher frequencies (Feijoo & Hernandez, 1990). The suggested formula that models this relationship is described as follows (Deller, Proakis & Hansen, 1993):…”

A screening test for speech pathology assessment using objective quality measures

“…The pitch perturbation measure implemented here is based on the work of Feijoo and Hern andez [6], which quanties how the pitch prole changes with time. The hypothesis is that a subject suering from a speech disorder will produce speech that is not as uid, strong, or constant as their healthy counterpart.…”

“…The amplitude perturbation (AP) measure [6] tracks the time evolution of a set of main amplitude peaks of the time-domain signal. The amplitude feature, denoted by SA, is determined by the maximum amplitude of the signal within a moving analysis window.…”

“…Historically, most objective speech quality measures have been formulated for assessing the performance of speech v ocoders [5]. More recently, several measures have been considered for assessment of severe pathology, such a s v ocal fold cancer [6]. This paper will describe a novel approach for integrating objective measures of speech quality with a proposed hidden Markov model (HMM) detection scheme for the purpose of developing a screening test capable of assessing subjects who are not producing high quality speech, a condition that may be indicative of a more serious speech disorder.…”