On Apr 8, 2011, at 2:08 PM, Hendrik Schreiber wrote: >> >> - There were 1-sample imprecisions in the matrices and kernel: for instance, >> a (symmetrical) kernel necessarily needs to have odd sizes of the form 2N+1, >> with N rows above the diagonal and N rows below, right? > > [...] > Therefore I believe that the even kernel is actually better suited (at least > for non-rotated kernels). Also, when looking at J. Foote, "Automatic Audio Segmentation using a Measure of Audio Novelty." In Proceedings of IEEE International Conference on Multimedia and Expo, vol. I, pp. 452-455, 2000. => http://rotorbrain.com/foote/papers/footeICME00.pdf Foote constructs even kernels. -hendrik