Toward the Quantification of Vocal Efficiency