Aspects of inverse filtering for loudspeakers


  • Scott G. Norcross Advanced Audio Systems, Communications Research Centre, Station H, 3701 Carling Ave., Ottawa, Ont. K2H 8S2


Discrete Fourier transforms, Frequency domain analysis, Impulse response, Least squares approximations, Optimization, Signal processing, Vectors, Wave filters, Audio quality, Frequency-domain deconvolution method, Inverse filtering, Linear filtering


The time-domain least squares approach and the frequency-domain deconvolution method for the calculation of inverse filters of a loudspeaker's IR are examined. The frequency-domain method is less robust than the time-domain method and regularization and complex smoothing help in improving the subjective performance of the inverse filter. The amount of regularization required is dependent on the IR that is to be inverted, and hence require hand-tuning for optimizing the subjective performance. The complex smoothing of the impulse response with a third octave smoother improves the subjective performance of the inverse filter and does not broaden the main pulse of the corrected IR.




How to Cite

Norcross SG. Aspects of inverse filtering for loudspeakers. Canadian Acoustics [Internet]. 2004 Sep. 1 [cited 2021 Dec. 5];32(3):156-7. Available from:



Proceedings of the Acoustics Week in Canada