Predicting speech intelligibility in rooms

Abstract

The pioneering work of Lochner and Berger [e.g., J. Sound Vib. 1(4), 426–454 (1964)] has provided a basis for measuring and calculating speech intelligibility in rooms in terms of a ratio of useful early sound to detrimental late arriving sound and background noise. In the present work their results are used in an image sources computer model to predict speech intelligibility in rooms. The calculation procedure is discussed and the influence on speech intelligibility of room geometry, room surface sound absorption properties, and background noise levels are investigated. Comparisons are made with speech intelligibility calculations based on the modulation transfer function approach.