Recent work in blind source separation applied to anechoic mixtures of speech allows for improved reconstruction of sources that rarely overlap in a time-frequency representation. While the assumption that speech mixtures do not overlap significantly in time-frequency is reasonable, music mixtures rarely meet this constraint, requiring new approaches. We introduce a method that uses spatial cue...

These notes were prepared for International Workshop on Harmonic and Quasiconformal Mappings in Chennai, August 2010. They are based on several papers written jointly with Tadeusz Iwaniec, Jani Onninen, and Ngin-Tee Koh. The source papers are available from www.arxiv.org

Introduction. Amidst the multiplicity of audio signal processing tasks connected with necessity source separation, problem obtaining its most prominent components occupies special place. One approaches, dedicated to solve such task, is based on melody extraction from musical recording. If generalized, technique can be applied a wider class signals for them so-called fundamental frequencies cont...