A New Approach For Detecting Strong Reflections In Rooms

Abstract: In this thesis the influence of strong reflections in rooms is investigated. A strong reflection can have a negative influence on speech or music that is transmitted from the source to the receiver. Impulse responses including different strength of a reflection are modelled using a statistical model. From the impulse response, different parameters describing the strength of the reflection can be calculated and the detection thresholds for the parameters are found from listening tests. Three different parameters were compared, of which one is an existing parameter by Dietsch and Kraak and the other two are new parameters based on the energy decay curve of the impulse response. The threshold is given as the minimum level at which the subject can hear the reflection by listening to the impulse response. The detection threshold is found from impulse responses for rooms with different volumes and reverberation times and for different delay times of the strong reflection. The performance of the parameters is then discussed and a suggested technique for identifying and evaluating a strong reflection is given.