Discovering Convolutive Speech Phones using Sparseness and Non-Negativity Constraints

Discovering a representation that allows auditory data to be parsimoniously represented is useful for many machine learning and signal processing tasks. Such a representation can be constructed by Non-negative Matrix Factorisation (NMF), which is a method for finding parts-based representations of non-negative data. Here, we present an extension to convolutive NMF that includes a sparseness constraint. In combination with a spectral magnitude transform of speech, this method extracts speech phones (and their associated sparse activation patterns), which we use in a supervised separation scheme for monophonic mixtures