This project is an investigation of the possibility of using deconvolution or time reverse acoustics to perform suppression of the reverberation of sound from public address systems in rooms. This is similar to the method by which we might suppress the feedback in conferencing telephone systems. However, in the case of room reverberation the challenge is that the impulse responses of the room are constantly changing (e.g. as people move in their seats). The project is to investigate the typical variability in room impulse responses from P/A loudspeakers over the rea of a seated audience and find if there might be techniques (e.g. frequency shifting) for suppressing some of the late-arriving sound to listeners in the reverberation excited by the sound system. Such an improvement in the direct-to-reverberant sound ratio received by listeners is useful in improving the intelligibility of speech from the PA system especially for hearing impaired listeners who comprise an increasing number of people in our aging population.
Undergraduate
An understanding of the statistics of room frequency responses and reverberation in rooms. How the intelligibility of PA systems is related to the room reverberation and the ratio of direct to reverberant sound received by listeners. How this ratio is affected by the directionality of the loudspeakers. Detailed measurements of the variation of room impulse responses over a defined seating area in a reverberant room and predictions and trials whether deconvolution of the loudspeaker input speech with a truncated generalized impulse response is of value in improving the direct to reverberant ratio. As time permits the value that frequency shifting of the loudspeaker output may have will also be investigated.
None
Acoustics Lab (City 422.154, Lab)