The authors would like to thank Beiming Wang, Harald Viste and Mathieu Lagrange for assistance with some of the figures in this chapter. MP was supported in part by EPSRC Leadership Fellowship EP/G007144/1 and by the EU Framework 7 FET-Open project FP7-ICT-225913-SMALL: Sparse Models, Algorithms and Learning for Large-Scale data.
1 Models (14.1) and (14.2) both assume that the source positions and the digital audio effects applied upon them are fixed throughout the duration of the recording. Moving source scenarios have received less attention so far. However, such scenarios may be addressed to a certain extent by partitioning the recording into time intervals over which the mixing filters are reasonably time-invariant and applying the techniques reviewed in this chapter to each interval.