| Fri, 25 Nov 2011 16:51:33 +0000 |
Andrew N Robertson |
limited loaded files to 10000 frames so that testing is easier. Great on classical, rubbish on rock and pop
|
| Wed, 23 Nov 2011 22:41:56 +0000 |
Andrew N Robertson |
fixed bug causing bad path for the first index. Added in chroma sim using Euclidean distance option as well as dot product
|
| Wed, 23 Nov 2011 18:33:17 +0000 |
Andrew N Robertson |
part optimised the calculation of combined energy and chrroma in partJointSim calc. Added bug fix
|
| Tue, 22 Nov 2011 17:37:45 +0000 |
Andrew N Robertson |
Causal version working with large files too, anchor points shown in white.
|
| Wed, 09 Nov 2011 18:34:25 +0000 |
Andrew N Robertson |
added get time now function with a view to allowing offline update too
|
| Mon, 07 Nov 2011 18:08:43 +0000 |
Andrew N Robertson |
fixed bug with switching and make sure alignIndex never goes to zero, added in limit to the joint matrix that gets created.
|
| Mon, 07 Nov 2011 17:24:52 +0000 |
Andrew N Robertson |
Version that works sequentially with chroma and onset energy, but doesn't use combined matrix, hence more efficient. playing switches okay, bug fixed on path calculation
|
| Fri, 04 Nov 2011 22:56:48 +0000 |
Andrew N Robertson |
bug with loading needs fixing - if you use fullfilename for second audio file, it wont work properly...
|
| Wed, 20 Jul 2011 12:06:45 +0100 |
Andrew N Robertson |
some work in improving the efficiency of doing chroma and onset analysis ahs been made - don't need to recalculate the chroma comparison each time
|
| Wed, 08 Jun 2011 17:35:56 +0100 |
Andrew N Robertson |
revised this to work with onset and chroma and looking now to do sequential DTW
|
