The purpose of this study is to examine the acoustic implications of using aspirated, well-coordinated, coup de la glotte, and hard glottal onset methods, in order to compare and contrast the radiated acoustic spectra. Twenty-five singers trained in bel canto singing style were asked to sing 5-second samples on three pre-determined pitches comprising the low, middle, and high range in male and female voices. Each participant was instructed and trained to sing the three pitches with the four methods. EGG was used with audio perception to verify onset type, and VoceVista Video Pro was used to analyze power spectra. A repeated measures multivariate analysis of variance (rMANOVA) was performed with the SPSS General Linear Model function, with onset type as the within-subjects variable to determine main effects and interaction effects on harmonic amplitude (up to 5000Hz) from the independent variables. A significant main effect was found for onset type and more specifically, a significant acoustic difference was found between the well-coordinated and coup de la glotte onsets. Substantial inconsistencies were found in the execution of the well-coordinated onset, as well as in participants' reported preferred onset compared to their baseline measurement of executed onset type. Intentional study of the phonatory onset beginning with adducted folds and low subglottal pressure may have important acoustic and efficiency consequences for bel canto singers. There is a need for specificity regarding the prephonatory glottal configuration in vocal pedagogy and voice science research, and a numerical classification system may provide a clearer picture in future onset investigations.