Title: Research Student
Research Lab: Centre for Digital Music
Research Topic: Sound source separation for music remixing
I am investigating time-frequency methods for audio source enhancement, suppression and filtering in spatial music signals (for example, stereo). Approximate W-disjoint orthogonality is one assumption on audio signals that can be used for this. It relies on the spatial configuration of the sources and assumes that each time-frequency point belongs exactly to one source. It has been used on separation of speech sources and can give good results. However, in its simple formulation it is not suitable for extracting music sources (for example, there is the problem of overlapping partials). Therefore, this project aims to develop techniques in a similar spirit which can be successfully applied to spatial music signals.
A. Nesbit, M. D. Plumbley and M. E. Davies. A signal-adaptive local cosine transform for source separation by time-frequency masking. To appear in: Proceedings of the ICA Research Network International Workshop (ICArn 2006), 18-19 Sept 2006.
A. Nesbit, M. E. Davies, M. D. Plumbley and M. B. Sandler. Source extraction from two-channel mixtures by joint cosine packet analysis. In: Proceedings of the 14th European Signal Processing Conference (EUSIPCO 2006), Florence, Italy, Sept 4-8, 2006.
M. D. Plumbley, S. A. Abdallah, T. Blumensath, M. G. Jafari, A. Nesbit, E. Vincent and B. Wang. Musical audio analysis using sparse representations. In COMPSTAT 2006 Proceedings in Computational Statistics, Rome, Italy, pp 104-117. Physical-Verlag, 2006.
A. Nesbit, L. Hollenberg and A. Senyard. Automatic transcription of Australian Aboriginal music. In Proceedings of the 5th International Conference of Music Information Retrieval (ISMIR 2004), pp 326-330, Barcelona, Spain, October 10-14, 2004.