Smoothened cepstral peak prominence in choral singers, trained singers and non-singers

Abstract

The aim of the study was to understand the vocal quality of voice and eect of speech tasks on Smoothened Cepstral peak prominence (Smoothened CPP) across singers. Three groups of participants com- prising of 16 choral singers, 13 trained singers and 16 non singers were considered. Speech-Tool program developed by Hillenbrand (1994) was used to analyze the Smoothened CPP. Mixed ANOVA was carried out to verify the statistical signicance of dierence across the groups and speech tasks. Both Choral singers and trained singers obtained higher smoothened cepstral peak prominence than non-singers. Within singers, trained singers obtained higher smoothened CPP values than Choral singers; however the dierence was not statistically signicant. With respect to speech tasks, smoothened CPP was found to be higher for sustained phonation than Reading and this dierence was found to be statistically signicant. Hence, Smoothened CPP was able to dierentiate singers and non-singers irrespective of the speech tasks used.