Objective: Sustained vowels are an important element of the auditory-perceptual and the acoustic assessment of voice quality. While many clinicians will use sustained /a/ for this purpose, some protocols also include other vowels. For patients, the difference between sustaining a vowel and singing a vowel can be blurry. The present study used the VoxPlot software to assess the effect of the vowel and the task (sustaining vs. singing) on 15 voice quality outcome measures.

Methods: Thirty-two professional female singers (average age 27.2 years) produced sustained vowels /a, i, u/. Three repetitions of each vowel were sustained in a speechlike fashion and as a sung middle C at 261 Hz. The recordings were analyzed with the VoxPlot software. Data were analyzed with repeated measures ANOVAs.

Results: There were significant main effects for the vowel for 10 out of the 15 voice quality measures in VoxPlot (CPPS, GNE, HNR, HNRd, HFNO, H1H2, Jitter Local, Jitter PPQ5, Spectral Slope, Spectral Tilt). The post hoc tests did not follow a systematic pattern. For the task (sustained vs. sung), significant main effects were found for 6 of the 15 VoxPlot measures (GNE, HNR, HNRd,Jitter PPQ5, Mean Pitch, PSD). For 5 out of the 11 VoxPlot measures with reported cutoff scores, some or all scores were outside the reported typical ranges (CPPS, GNE, HFNo, HNR, PSD).

Conclusions: Most of the voice quality measures in VoxPlot differed for the three vowels but the post hoc patterns of differences were not consistent. When using software for acoustic analysis such as VoxPlot, it is important for clinicians to understand that the selection of the vowel may influence the results. There were significant task effects for some outcome measures, so clinicians should understand that they may get different results if they let their patients sing sustained vowels. Future research should explore whether differences between sung and spoken vowels are more pronounced in speakers with dysphonic voices.