Essentially, all models are wrong, but some are useful - George E.P. Box
The problem of acoustic source localization is important in many acoustic signal processing applications, such as distant speech acquisition and automated camera steering. In noisy and reverberant environments, the source localization problem becomes challenging and many existing algorithms deteriorate. Three-dimensional source localization presents advantages for certain applications such as beamforming, where we can steer a beam to both the desired azimuth and the desired elevation. In this paper, we present an acoustic source localization method with low computational complexity which, instead of using individual microphone signals, combines them to form eigenbeams. We then use the zero- and first-order eigenbeams to compute a pseudointensity vector pointing in the direction of the sound source. In an experimental study, the proposed method’s localization accuracy is compared with that of a steered response power localization method, which uses the same eigenbeams. The results demonstrate that the proposed method has higher localization accuracy.
Published in the Proc. of the 18th European Signal Processing Conference (EUSIPCO 2010), Aalborg, Denmark, Aug. 23-27, 2010.
[Back]