Previous work has shown that various flavours of Independent Component Analysis, when applied to natural images, all result in broadly similar localised, oriented band-pass feature detectors, which have been likened to wavelets or edge detectors. In this paper, we present a similar analysis of 'natural' sounds drawn from two radio stations: one broadcasting mainly speech; the other mainly classical music. Many of the resulting basis vectors are quite wavelet-like, and can easily be characterised in terms of their position and spread in the time-frequency plane. Some of them, however, particularly from the set trained on music, do not fit that interpretation very well. The Wigner-Ville Distribution can be used to gain a clearer picture of time-frequency localisation of these basis vectors. We conclude by suggesting that these results be compared with other widely used auditory representations such as short-term Fourier transforms, wavelet transforms, and physiologically derived models based on the auditory filter-bank.
展开▼
