The effects of microphone positioning in hearables on voice quality and F0 measurements.

Abstract

Voice quality and fundamental frequency (F0) metrics are important indicators of motor function and hold promise for health monitoring. Recent advances in hearables have enabled the longitudinal monitoring of speech production and its changes. Hearables can record speech from in-ear microphones (IEMs) and outer-ear microphones (OEMs), but it remains unclear how these measurements from hearables compare to the laboratory gold standard, a microphone placed in front of the mouth. This study examines voice quality and F0 measurements across the IEM, OEM, and the standard method (REF) using parallel recordings. Results showed that the IEM introduced more variability overall; increases in jitter, harmonic-to-noise ratio (HNR), F0 maximum, and standard deviation and decreases in F0 minimum were seen for females. Decreased shimmer and increased HNR were seen in the OEM. The causes of these differences were discussed. The findings indicate that the hearable-based measurements may not align with REF standards, suggesting the need for new standards specific to hearables. Preliminary observations of sex-based differences require further investigation with adequately powered and balanced samples to determine their significance and generalizability. Future research should further explore factors such as occlusion effect and sex-specific differences (e.g., F0 range) in the relationship between hearables and REF measurements.