Samia Fraj

Validity of jitter measures in non-quasi-periodic voices. Part II: The effect of noise

Abstract In this paper the effect of noise on both perceptual and automatic evaluation of the glottal cycle length in irregular voice signals (sustained vowels) is studied. The reliability of four tools for voice analysis (MDVP, Praat, AMPEX, and BioVoice) is compared to visual inspection made by trained clinicians using two measures of voice signal irregularity: the jitter (J) and the coefficient of variation of the fundamental frequency (F0CV). The purpose is also to test to what extent of irregularity trained raters are capable of determining visually the glottal cycle length as compared to dedicated software tools. For a perfect control of the amount of jitter and noise put in, data consist of synthesized sustained vowels corrupted by increasing jitter and noise. Both jitter and noise can be varied to the desired extent according to built-in functions. All the tools give almost reliable measurements up to 15% of jitter, for low or moderate noise, while only few of them are reliable for higher jitter and noise levels and would thus be suited for perturbation measures in strongly irregular voice signals. As shown in Part I of this work, for low noise levels the results obtained by visual inspection from expert raters are comparable or better than those obtained with the tools presented here, at the expense of a larger amount of time devoted to searching visually for the glottal cycle lengths in the signal waveform. In this paper it is shown that results rapidly deteriorate with increasing noise. Hence, the use of a robust tool for voice analysis can give valid support to clinicians in term of reliability, reproducibility of results, and time-saving.

Development and perceptual assessment of a synthesizer of disordered voices.

A synthesizer is based on a nonlinear wave-shaping model of the glottal area, an algebraic model of the glottal aerodynamics as well as concatenated-tube models of the trachea and vocal tract. Voice disorders are simulated by way of models of vocal frequency jitter and tremor, vocal amplitude shimmer and tremor, as well as pulsatile additive noise. Six experiments have been carried out to assess the synthesizer perceptually. Three experiments involve the perceptual categorization of male synthetic and human stimuli and one the auditory discrimination between synthetic and human tokens. A fifth experiment reports the auditory discrimination between synthetic tokens with different levels of additive and modulation noise. A sixth experiment reports the scoring by expert listeners of male synthetic stimuli on equal-appearing interval scales grade-roughness-breathiness (GRB). A first objective is to demonstrate the ability of the synthesizer to simulate vowel sounds that are valid exemplars of speech sounds produced by humans with voice disorders. A second objective is to learn how human expert raters perceptually map vocal frequency, additive and modulation noise as well as vowel categories into scores on GRB scales.

Validity of jitter measures in non-quasi-periodic voices. Part I: Perceptual and computer performances in cycle pattern recognition

Abstract The limit of about 5% for reliable quantification of jitter in sustained vowels of dysphonic voices—a widely accepted guideline—deserves critical analysis. The present study pertains to the effect of experience and training on the perceptual (visual) capability of correctly identifying periods in (highly) perturbed signals, and to a comparison of the performance of several programs for voice analysis. Synthesized realistic vowels (/a:/) with exactly known jitter (2.7%–31.5%) are used as material. After selection and training, experienced raters demonstrate excellent agreement in correctly identifying periods up to high values of jitter put in. Perceptual rating outperforms all computer programs in accuracy. Most remain reliable up to 10% jitter; one of them correctly measures up to the highest level.

To what degree of voice perturbation are jitter measurements valid? A novel approach with synthesized vowels and visuo-perceptual pattern recognition

Abstract Objective measurement of the severity of dysphonia typically requires signal processing algorithms applied to acoustic recordings. Since Lieberman (1963) introduced the concept of perturbation analysis in the area of voice, the best-known acoustic parameter in clinical practice is conventional jitter . However, jitter measurements have some critical limitations. According to a widely accepted guideline, in sustained vowels of dysphonic voices, only perturbation measures less than about 5% are reliable: this is related to period extraction methods. This limit of 5% deserves critical analysis, certainly when there are indications that some acoustic analysis programs can be applied to quite irregular voices such as substitution voices. The present experiment demonstrates that – on signals of synthesized deviant voices (sustained vowel) with moderate additive noise – different raters are able to visually identify in a very consistent way the period durations of successive cycles up to values of about 13% jitter. Furthermore, even for higher values – over 30% – the jitter % computed with the period values rated by visual perception is, for some of the raters, very comparable to the real value. This suggests that improved acoustic programs using more reliable algorithms could validly transgress the traditional limit of 5% if they demonstrate the correspondence of their computations with the true jitter values. This is now made possible by synthesizers generating artificial deviant voices that cannot be distinguished from true dysphonia, and in which the jitter put in is exactly known.

Testing the reliability of Grade, Roughness and Breathiness scores by means of synthetic speech stimuli

Abstract This article describes a synthesizer of disordered voices and reports a test of the reliability of Grade, Roughness, and Breathiness scores assigned to synthetic stimuli by eight expert listeners in two sessions. Speech stimuli [a], [i], [u], [ai], and [ia] were synthesized with three values of vocal frequency and four levels of vocal jitter and pulsatile additive noise each. The agreement and correlation of scores assigned by the same rater in different sessions, or by different raters in the same session, accord with published data. Only a small part of the variance of the arithmetic differences between the scores that are assigned to the same stimulus is explained by the stimuli properties. The conclusion is that differences between scores that are assigned to the same stimulus are not attributable to biases of individual raters; such biases would shift all the scores assigned on a scale, and the shift would be interpretable in terms of the properties of the stimuli.