Influence of Measurement Procedure on the Korean Matrix Sentence-in-Noise Intelligibility for Normal-Hearing Listeners

Abstract

Background and Objectives A Korean Matrix Sentence Test is one of the adaptive tests used to evaluate speech intelligibility in noise. The present study aimed to investigate the influence of measurement procedure (presentation level and mode) on the Korean matrix sentence-in-noise intelligibility of normal-hearing listeners.Subjects and Method Twenty-seven normal-hearing adults participated in the study. The speech reception threshold (SRT) was determined using a software-based adaptive procedure, which converges on the signal-to-noise ratio (SNR), corresponding to 50% intelligibility. The speech or speech-shaped noise was given at three presentation levels based on the previous categorical loudness judgments (between soft and medium: 55 dB SPL, medium: 65 dB SPL, and loud level: 80 dB SPL). The effect of the presentation mode was explored by fixing either the sentence level or noise level, while the noise or sentence level was adaptively adjusted according to the subject’s response.Results The presentation level significantly affected the SRTs, whereas the SRTs did not depend on the presentation mode. The sentence-in-noise intelligibility was similar between 55 and 65 dB SPL, yet the sentence-in-noise intelligibility significantly decreased at 80 dB SPL regardless of the presentation mode. Correlation analyses showed moderate to high positive correlations among the results of two presentation modes.Conclusion A presentation level of 55 or 65 dB SPL yielded comparable adaptive measurement results for normal-hearing subjects, yet the sentence-in-nose intelligibility became worse at 80 dB SPL. Clinicians should be careful with the influence of the presentation level when identifying speech-in-noise intelligibility.