This paper deals with Cepstral Peak Prominence Smoothed (CPPS) distribution and its descriptive statistics as possible indicators of vocal health status. 41 voluntary patients and 35 control subjects participated in the experiment: all of them followed the same protocol, which includes three repetitions of the sustained vowel /a/ simultaneously acquired with a microphone in air and a contact sensor, the perceptual assessment of voice quality and the videolaringoscopy examination. The fifth percentile and the standard deviation of CPPS distribution were the parameters included in the best logistic regression models for the microphone in air and the contact sensor, respectively. The selected CPPS parameters had a strong to good discrimination power: an Area Under Curve of 0.95 and 0.87 has been found for the microphone in air and for the contact sensor, respectively. For each CPPS parameter, the repeatability has been also estimated and the Monte Carlo method has been implemented for the uncertainty evaluation of the discrimination threshold. Furthermore, preliminary recommendations for better accuracy and repeatability of future studies are provided: analyses on the main CPPS influence quantities and on the effect of the frequency content of the signal spectrum on the CPPS parameters have been provided. Index TermsCepstral analyses, human voice, biomedical measurement, acoustic devices, reproducibility of results, Monte Carlo methods, uncertainty I. INTRODUCTION Traditionally, voice quality has been assesed using subjective tests, in which experts listen to live or recorded vocal signals and perceptually rate them. In order to overcome the subjectivity and the expensiveness of such methods and with the aim to find a less time-consuming tool, researchers started to analyze voice signals and to extract several parameters as indexes of different aspects of voice and voice-related issues.

Recent literature reports that a large percentage of teachers complain that teaching has an adverse effect on their voice status. Thus, more needs to be done to study their vocal behavior. The objective of this longitudinal study was twofold: to determine changes in the voice use of teachers over a school year, and to study the relationships between voice use and classroom acoustic parameters. Thirty-one teachers from two secondary schools in Turin (Italy) were involved at the beginning of the 2014-2015 school year, and 22 of them also participated at the end of the same school year. The results show that teachers adjust their voices with noise and reverberation. A minimum value of the sound pressure level of voice (SPL) was found at a mid-frequency reverberation time of 0.8 s in both periods. Moreover, the teachers who worked in the worst classroom acoustic conditions showed an increase of 2.3 dB in the mean SPL and a decrease of 10% in the voicing time percentage at the end of the school year. A predictive model that can be used to estimate the mean SPL from the background noise level and the reverberation time, based on collected data, is here proposed.

No abstract

The intra- and inter-speaker variability of speech sound pressure level (SPL) has been investigated under repeatability conditions in this work. In a semi-anechoic chamber, speech from 17 individuals was recorded with a sound level meter, a headworn microphone, and a vocal monitoring device. The subjects were asked to read twice and in sequence two phonetically balanced passages. The speech variability has been investigated for mean, equivalent, and mode SPL from each reading and device. The intra-speaker variability has been evaluated by means of the average among individual standard deviations in the four readings and it reached the maximum of 2 dB for mode SPL. For the inter-speaker variability, the experimental standard deviation of individual averaged SPL parameters among the four repeated measures has been calculated, obtaining the highest value of 5.3 dB for mode SPL. Changes in SPL variability have been evaluated with different logging intervals for each device. The influence of speech material has been investigated by the Wilcoxon test on paired lists of descriptive statistics for SPL distribution and equivalent SPL in the repeated readings. The data reported in this study may be considered as a preliminary reference for the investigation of changes in speech SPL over subjects.

No abstract