The Acoustic Voice Quality Index (AVQI) in the assessment of voice quality: A preliminary multiple-case study of the effects of recording level, room noise and phonation type

Authors

• Anne-Maria Laukkanen Speech and Voice Research Laboratory, Faculty of Social Sciences, Tampere university
• Tero Ikävalko Speech and Voice Research Laboratory, Faculty of Social Sciences, Tampere university
• Leena Rantala Speech and Voice Research Laboratory, Faculty of Social Sciences, Tampere University / Unit of Logopedics, Faculty of Social Sciences, Tampere University
• Eliina Kankare Phoniatric clinic, Tampere University Hospital

DOI: https://doi.org/10.23997/pk.101515

Keywords:

acoustic analysis, dysphonic voices, recording, room acoustics

Abstract

The AVQI is used for the evaluation of dysphonia. This preliminary study investigated how
much the index is affected by a low input gain level or background noise during recording,
as well as the speaker’s gender and way of speaking when the recommended recording device
are used.

Samples were obtained from two vocally healthy speakers and two voice patients (male and
female in both cases). The healthy participants were recorded in an acoustically treated studio
with a head-set microphone while reading a passage and sustaining [a:] in a habitual way,
softly, loudly, and in a higher than the habitual pitch. The recordings and the voice patients’
samples were re-recorded in a studio and in an ordinary office using an optimal and low gain
level. The AVQI was calculated.

The recording level, room noise, and speaker’s gender had a minor effect on the AVQI. The
dysphonic voices obtained the highest (= worst) AVQI values. The healthy participants’ soft
samples resulted in AVQI values typical of dysphonia. When the AVQI is utilized for voice
evaluation, it is important to ensure that the voice samples are produced at a conversational
loudness.