To confirm the accuracy of the algorithm for determining vocal dose and effort, four singers (soprano, alto, tenor, and baritone) completed a set of between 10 and 11 tasks while recording with the MA device, each lasting for 1 min with 1 min of rest in between. A handy video recorder (Q8, Zoom) for all participants simultaneously recorded acoustic audio samples. The tasks included normal speaking, speaking over 60 dB of ambient noise, whispered speaking, moderately loud singing (low to mid range), moderately loud singing (mid to high range), very loud singing (low to mid range), very loud singing (mid to high range), singing without vibrato (low to mid range), staccato arpeggios throughout the vocal range, and strained speaking. Each participant selected one excerpt to use for the low- to mid-range tasks and another piece to use for the mid- to high-range tasks. For the speaking exercises, participants read from “Practicing Vocal Music Efficiently and Effectively: Applying ‘Deliberate Practice to a New Piece of Music’” by Ruth Rainero (21 ). The strained speaking task was completed only by the alto participant.