Abstract
A method for determination of the transmission coefficient from Complex Modulation Transfer Functions CMTF:s based on measured impulse-responses is shown.
In the method a separate measurement of the equivalent sound absorption area is not needed in contrast to in the standardised measurement.
By averaging over a number of estimates of the impulse-response the influence of background noise can be reduced substantially, implying that low-power sources can be used.
A model for the power balance in the receiving room with time-varying power is considered. In the model the quotient of the receiving room intensity and the source room intensity has one pole, which is proportional to the equivalent sound absorption area in the receiving room, and a gain, proportional to the transmission coefficient.
In the physical system the power can be time-varied by letting the system excitation signal consist of random noise modulated with a deterministic time-varying function. However, since the ensemble average of the squared response is proportional to the squared impulse-response convolved with the squared modulating function, random excitation is avoided and replaced by impulse-response measurements. The quotient of intensities in the model is in the physical system a quotient of CMTF:s.
Experiments are carried out in an airborne sound insulation laboratory. For comparison, standardised measurements are also carried out. It is found that the presented method gives as result comparatively small transmission coefficients, though the relative differences are small. By refining the power balance model by introducing an energy propagation time delay, and selecting an appropriate delay, the differences were diminished.