Speech Intelligibility Quality Research Articles

Acoustic transfer functions and speech intelligibility in permanent wearing headsets

A detailed analysis of the acoustic transfer functions has been investigated for different locations of the headset microphones in the area of human head. Amplitude-frequency and phase-frequency characteristics of acoustic signals are discussed in the report. The signals have been recorded with the micro-mechanical-electrical-system (MEMS) microphones located on a mannequin, simulating the acoustic properties of the human torso. The the source of the signals is an emitter built in the mouth of the manikin's head. Analysis of the characteristics has shown that the acoustic signals undergo significant phase and amplitude changes associated with the interference and diffraction processes. These changes result in significant nonlinearity phase characteristics, and this nonlinearity depends on the installation location of the MEMS microphone. Also quality of speech intelligibility has been estimated by the method of expert evaluation and the speech has been recorded for different positions of the microphones and speaker head. A comparative analysis of the results has shown the optimal location of the microphones in the ear and on the breast of man.

Relations between rapid speech transmission index (RASTI) and other acoustical and architectural measures in churches

The use of rapid speech transmission index (RASTI) in churches is studied and the relationships with acoustical and architectural parameters identified using a major survey of Catholic churches in Portugal. It was found that the vast majority of churches tested have RASTI values below 0.45 giving a poor rating in the quality of speech intelligibility. RASTI values within churches can be reasonably predicted by the use of center time (TS) at 1000 Hz. Loudness (L) does not appear as an important characteristic regarding RASTI values supporting the idea that the intelligibility of speech, under reverberant conditions does not depend on loudness. A prediction equation using three architectural parameters was calculated to estimate the average RASTI in churches. The effect of the architectural styles and the use of pulpits on RASTI values was also studied.