Measured room impulse responses (RIR) strongly depend on the directivity of the sound source used for the measurement. An analysis method is presented that is capable of pinpointing the influence of the loudspeaker's directivity on a set of RIRs. Taking into account the rotational symmetries of a dodecahedron loudspeaker, it detects the effects that the changing directional pattern induces in the RIR. The analysis of RIRs measured in completely different acoustical environments reveals that the influence of the loudspeaker's directivity can still be observed in the very late part of the RIR-even in very reverberant rooms. These results are presented and the consistency with general room acoustical theory is revised and discussed.

In order to improve the understanding of uncertainties in measuring the acoustics in auditoria, the influence of a sound source's directivity is investigated. In previous work a model to predict the uncertainties when measuring room impulse responses with sources of a given directivity pattern has been developed. As a result, properties of the measurement environment, i.e. the size of the room, its reverberation as well as the sound scattering behaviour of the room surfaces, were identified to be significant secondary influences. Through extensive series of scale measurements data has been collected in a reverberation room to validate the model prediction. By introducing adjustable partition panels, absorbing and sound scattering surface elements the secondary influence factors were carefully controlled over a large range of values. After a brief explanation of the uncertainty model the results of the validation measurements will be presented. The significance of the different influence factors on the measurement uncertainty will be discussed.

In previous investigations, it has been shown that the directivity of a measurement sound source has a significant influence on the measured room impulse response (RIR). Using a specialized method of analysis, the sources influence can be identified even in the very late part of the RIR even in very reverberant environments. These results seem to be surprising at first and contradict intuitive expectations. In this contribution, the findings are briefly discussed, and the congruence with general room acoustic theory is revised and discussed.

No abstract