Validation of the Acoustic Voice Quality Index version 03.01 in the French language

Summary: Objectives. The Acoustic Voice Quality Index (AVQI) version 03.01 is a tool for quantitative assessment of the overall severity of dysphonia. Its computation includes six acoustic parameters, which are all carried out using the freeware Praat. It is based on the recordings of a sustained vowel and a part of a text read aloud. In order to obtain a time balance between the sustained vowel and the continuous speech, Barsties and Maryn (2015) have found that — for the Dutch sample — 34 syllables have to be read aloud. The aim of this preliminary study was to determine how many syllables of a commonly used text in the French-speaking part of Belgium should be used to reach the same time balance.
Methods. In order to identify the optimal part of the text to be taken into account for the AVQI’s computation, we first looked for the phonemic contents of different extracts in comparison to the phonemic frequencies in the French language. We then confirmed our choice on a larger sample from the database of the University Hospital of Liège (n=162). Therefore, we manually measured the duration of each recording’s concatenated voiced segments. The descriptive data of these lengths was then used to confirm the optimal length of the text extract to be used. This extract allows to obtain a time balance of the running speech sample with the central three seconds of the sustained vowel.
Results. 27 syllables were found to be the most appropriate duration of the text to be read aloud, in order to balance out the time analysed for both the sustained vowel and the continuous speech. The median duration of the concatenated voiced segments for the 162 recordings is three seconds (interquartile range: 0.76 seconds), which is well-balanced with the three seconds of the sustained vowel.