Prediction of classroom speech clarity using backpropagation neural network

Abstract

A classroom's acoustic design should be constructed so that the highest possible degree of speech intelligibility is achieved for teachers and students. In achieving the highest possible speech intelligibility, the acoustical design of classrooms should be based on the listener's positions in the classroom. The variation of Speech Clarity (SC) in a University Classroom with respect to Signal to Noise Ratio (SN) is dealt in this paper. The speech levels are measured at different seating position and the SC is determined. A simple Neural Network model was developed to predict the SC at any point in the classroom.