Noise-Robust Monitoring of Lombard Speech Using a Wireless Neck-surface Accelerometer and Microphone
Abstract
Ambulatory monitoring of voice characteristics has the potential to provide important data for the assessment of voice and speech disorders and psychological and emotional state. In this paper, we report on the development of a lightweight, wireless voice monitor that synchronously records dual-channel data from an acoustic microphone and a neck-surface accelerometer. Pilot data were collected from four adult speakers with normal voices who read aloud a phonetically balanced paragraph in the presence of increasing background acoustic noise levels to evaluate the signal-to-noise ratio (SNR) of both sensors and quantify Lombard speech effects. As expected, the SNR of the non-acoustic accelerometer sensor remained high in the presence of loud background noise levels and was more robust than that of the microphone. Lombard speech was observed by all four speakers who exhibited increases in accelerometer-based estimates of mean sound pressure level (+2.3 dB), fundamental frequency (+21.4 Hz), and cepstral peak prominence (+1.3 dB) in the presence of elevated background noise levels. Future work calls for ambulatory data collection using the wireless voice monitor in naturalistic environments with a larger cohort of speakers with various voice disorders, neurological conditions, and cognitive load levels.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 20, 2017
- Accession Number
- AD1032776
Entities
People
- Daryush D. Mehta
- Laura J. Brattain
- Patrick C Chwalek
- Thomas F. Quatieri
Organizations
- MIT Lincoln Laboratory