| Abstract | Objective: This study aimed to explore the correlation between the perceptual pleasantness/ring in the voice and its corresponding quantitative acoustic parameters, specifically Singer’s Formant (Fs) and Singing Power Ratio (SPR). The objective of the study was to examine the correlations between auditory-perceptual judgment of resonant voice, Fs, and SPR post-training.
Method: Twenty-nine trained Carnatic classical singers (18 females and 11 males) in the age range of 20 to 35 years (M=22; S.D=2.09) underwent 21 days of Resonant Voice Therapy. Baseline sung samples were recorded. After the training, sung samples were again obtained from all the participants. The singing samples were subjected to Fast Fourier Transform analysis for extraction of Long-Term Average Spectrum (LTAS). The amplitude of the Fs was manually noted from the LTAS, and the SPR was calculated by subtracting the amplitude of the highest peak between 2-4kHz from the highest amplitude in the region of 0-2kHz. The perceptual judgment of the vocal resonance/pleasantness of voice was checked using an 11-point equal appearing interval (EAI) scale by five qualified Speech Language Pathologists who were also trained Indian classical singers. A Spearman’s correlation coefficient was employed to analyze the correlations between perceptual judgment, amplitude of Fs, and SPR values.
Results: Twenty-four singers were observed to have Singers’ formant with a median amplitude of 27.69(males) and 27.75(females) participants. SPR median values were -23.09 (males) and -19.38 (females). Perceptual ratings showed a median of 5.6 for male participants and 6.2 for females. Spearman’s correlation indicated strong to moderate significant correlations between Fs amplitude and SPR values. However, perceptual judgments had moderate to weak, insignificant correlations with both Fs amplitude and SPR values.
Conclusions: The results showed that a strong correlation existed between Fs amplitude and SPR values, with Fs showing a better correlation with perceptual judgments than SPR. The parameter of SPR could be extracted from all voices, whereas Fs could be absent in some voices. These findings call for further studies to delve into the correlation between perceptual attributes and acoustic measures.
|
|---|