Speech Intelligibility: Lateral Lisp Detection for Real-Time Monitoring

Abstract Correcting lisps in speech can prove to be of great difficulty many, as they may unaware whether are lisping. To help those affected, we have developed a simple algorithm for the real-time identification sigmatismus lateralis ”S” sounds within via analysis frequency domain. The identifies peaks lisp’s band after calibration. A 3000-4000 Hz has been identified generally accurate lisp and 2500-3000 correct pronunciation single male test subject. splits given recordings into smaller segments compares number non-lisps detected these categorize. From tests, it was concluded that segment length 0.5 s produces best results. does not detect every section, however raise false positives. Our implementation Julia with multi-threaded per-file is able analyze 20 files lengths between 5 10 0.21 on Qualcomm Snapdragon 860 smartphone chipset, meaning far faster than real-time. proposed prototype capable realtime audio domain identify lateral dominant sibilant window. method only tested However, calibration adjusting parameters new individuals proposed. itself should easily expandable other impediments.