Characterizing vocal-fold dynamics in singing vocal modes from Complete Vocal Technique using high-speed laryngeal imaging and electroglottographic analysis

Mari Leppävuori (1) , Gustavo Andrade-Miranda (2) , Nathalie Henrich Bernardoni (3) , Anne-Maria Laukkanen (4) , Ahmed Geneid (5, 6)

1 University of Oulu
2 UG - Universidad de Guayaquil
3 GIPSA-VSLD - GIPSA - Voix Systèmes Linguistiques et Dialectologie
4 TUT - Tampere University of Technology [Tampere]
5 HUS - Helsinki University Hospital [Finland]
6 Helsingin yliopisto = Helsingfors universitet = University of Helsinki

Complete Vocal Technique (CVT) is a method used in voice pedagogy. In CVT, all voiced sounds (with vowels) are divided into four vocal modes – Neutral, Curbing, Overdrive and Edge. The vocal modes are identified and defined according to auditory attributes - metal, density, vowel and loudness – setting practical rules for the use of vocal modes. Recent laryngostroboscopic imaging studies have demonstrated visibly different laryngeal gestures between Curbing, Overdrive and Edge (McGlashan et al., 2016; Thuesen et al., 2017; Aaen et al., 2018). There are no study on vocal-fold dynamics of the vocal modes using high-speed imaging.The aim was to examine vocal-fold dynamics in CVT four vocal modes by means of laryngeal imaging techniques and electroglottography (EGG) analyses. Furthermore, the purpose was to test the feasibility of the selected methods to assess CVT vocal modes.Five non-classical singers (3 F mean age 38 yo; 2 M mean age 41 yo) trained in CVT were imaged with highspeed endoscopic camera while singing sustained vowels at the same pitch (F4 for male and Bb4 for female) in all vocal modes. Audio and electroglottography signals were simultaneously recorded. For each vocal mode, the singer could repeat the task several times in order to produce the most representative samples for further analysis. Stable parts of the samples were selected using Praat software. Audio and EGG signals were processed using Matlab software. Highspeed images were processed using kymography and optical-flow analyses (Andrade Miranda et al., 2017).The results show that contact duration, as evidence by EGG-derived contact quotient CQ, is greater than 44% of glottal-cycle duration for the three modes Curbing, Overdrive, and Edge. Lowest values and greater variability are found in Neutral mode, with contact quotient ranging between 12-57% of glottal-cycle duration. These vibratory behaviors are reflected on high-speed kymograms and optical-flow playbacks.Concerning the feasibility of highspeed endoscopy for vocal-mode assessment, singers had difficulty to produce vocal modes based on closed vowels, such as Curbing. The rigid endoscopy interferes with the singer’s lingual and labial articulatory behavior modifying the resulting vowel and the expected sound for a given vocal mode.

Mots clés: complete vocal technique; electroglottography; highspeed imaging; singing

Domaines: Linguistique; Acoustique [physics.class-ph]; Traitement du signal et de l'image [eess.SP]; Musique, musicologie et arts de la scène

Mari Leppävuori, Gustavo Andrade-Miranda, Nathalie Henrich Bernardoni, Anne-Maria Laukkanen, Ahmed Geneid. Characterizing vocal-fold dynamics in singing vocal modes from Complete Vocal Technique using high-speed laryngeal imaging and electroglottographic analysis. PEVOC 2019 - 13th Pan-European Voice Conference, Aug 2019, Copenhagen, Denmark. ⟨hal-02301637⟩