| Sat, 01 Jun 2013 22:05:01 +0100 |
Chris Cannam |
Make convolver tail-recursive
|
| Sat, 01 Jun 2013 19:20:19 +0100 |
Chris Cannam |
Make convolver options into options
|
| Sat, 01 Jun 2013 18:53:46 +0100 |
Chris Cannam |
Fix overlap-add in convolver
|
| Sat, 01 Jun 2013 13:21:28 +0100 |
Chris Cannam |
Fast convolver now works, up to some error -- though still only mono and with a fixed convolution framesize
|
| Sat, 01 Jun 2013 13:20:46 +0100 |
Chris Cannam |
Fix amazing error in complex (and add test for it)
|
| Sat, 01 Jun 2013 12:21:15 +0100 |
Chris Cannam |
Fast convolution runs, returning wrong results
|
| Fri, 31 May 2013 22:04:27 +0100 |
Chris Cannam |
Plug in fast convolver (doesn't work yet though!)
|
| Fri, 31 May 2013 22:04:05 +0100 |
Chris Cannam |
Make realInverse able to take a list
|
| Fri, 31 May 2013 22:03:45 +0100 |
Chris Cannam |
Fix complex test path
|
| Fri, 31 May 2013 22:03:23 +0100 |
Chris Cannam |
Replace add c1 c2 with sum [c1, c2]
|
| Fri, 31 May 2013 22:02:54 +0100 |
Chris Cannam |
Supply list to vector add
|
| Fri, 31 May 2013 22:02:24 +0100 |
Chris Cannam |
Make add take a list rather than just 2
|
| Fri, 31 May 2013 22:02:03 +0100 |
Chris Cannam |
Move complex tests to complex package
|
| Fri, 31 May 2013 17:34:46 +0100 |
Chris Cannam |
Make a start on fast convolver
|
| Fri, 31 May 2013 17:34:37 +0100 |
Chris Cannam |
Remove dependency from framer on filter (avoiding circular dependency)
|
| Fri, 31 May 2013 17:34:04 +0100 |
Chris Cannam |
Add entrywise product
|
| Fri, 31 May 2013 17:32:56 +0100 |
Chris Cannam |
Make boxcar all-1s, not all-0.5s -- means it doesn't have the same attenuation on average as other windows, but it's more consistent with other environments
|
| Fri, 31 May 2013 15:14:09 +0100 |
Chris Cannam |
Hurrah! Overlap-add now produces valid output streams (finally) and passes the tests
|
| Fri, 31 May 2013 10:58:00 +0100 |
Chris Cannam |
Revise overlap-add
|
| Thu, 30 May 2013 22:33:16 +0100 |
Chris Cannam |
More overlap-add tests.
|
| Thu, 30 May 2013 20:27:54 +0100 |
Chris Cannam |
Window overlap scale factor
|
| Thu, 30 May 2013 17:56:29 +0100 |
Chris Cannam |
Periodic window probably better after all
|
| Thu, 30 May 2013 17:47:25 +0100 |
Chris Cannam |
Further test
|
| Thu, 30 May 2013 17:34:35 +0100 |
Chris Cannam |
More on overlap-add for frames
|
| Wed, 29 May 2013 22:59:23 +0100 |
Chris Cannam |
Toward frame -> stream processing. Need overlap-add.
|
| Wed, 29 May 2013 17:32:58 +0100 |
Chris Cannam |
Merge
|
| Wed, 29 May 2013 17:32:27 +0100 |
Chris Cannam |
Start on un-framing (does not build)
|
| Fri, 24 May 2013 12:41:41 +0100 |
Chris Cannam |
Synchronise audio read
|
| Tue, 28 May 2013 20:41:11 +0100 |
Chris Cannam |
Make convolver work. Ugly though
|
| Mon, 27 May 2013 23:37:57 +0100 |
Chris Cannam |
Simplistic convolution implementation and (currently failing) tests
|
| Sat, 25 May 2013 18:42:01 +0100 |
Chris Cannam |
Fixes to duplicate
|
| Sat, 25 May 2013 18:41:53 +0100 |
Chris Cannam |
(clumsily) make slice functions handle out-of-range arguments properly
|
| Sat, 25 May 2013 18:31:10 +0100 |
Chris Cannam |
Make slice return partial slices when range extents overlap vector ends
|
| Sat, 25 May 2013 18:18:10 +0100 |
Chris Cannam |
Implementation and (currently) failing tests for duplicate
|
| Thu, 23 May 2013 21:36:20 +0100 |
Chris Cannam |
Remove obsolete bits
|
| Thu, 23 May 2013 19:33:06 +0100 |
Chris Cannam |
Move some of the most commonly loaded modules up a level
|
| Thu, 23 May 2013 17:15:27 +0100 |
Chris Cannam |
Quicker, though uglier, sparse products
|
| Thu, 23 May 2013 16:12:21 +0100 |
Chris Cannam |
Faster sparseProductLeft. Similar code would work for other sparseProducts
|
| Thu, 23 May 2013 14:45:21 +0100 |
Chris Cannam |
Add note for docs
|
| Thu, 23 May 2013 14:39:40 +0100 |
Chris Cannam |
Add repeated
|
| Thu, 23 May 2013 14:39:18 +0100 |
Chris Cannam |
Add vector plot type
|
| Thu, 23 May 2013 14:38:57 +0100 |
Chris Cannam |
Simpler audiofile test with more slack
|
| Thu, 23 May 2013 13:34:27 +0100 |
Chris Cannam |
Further tests and fixes
|
| Thu, 23 May 2013 13:21:05 +0100 |
Chris Cannam |
Window fixes and tests
|
| Thu, 23 May 2013 11:21:30 +0100 |
Chris Cannam |
transform/window -> signal/window
|
| Thu, 23 May 2013 11:20:15 +0100 |
Chris Cannam |
Window functions: make symmetric/periodic an option
|
| Wed, 22 May 2013 22:19:48 +0100 |
Chris Cannam |
Fix window functions. We need tests for these.
|
| Wed, 22 May 2013 15:02:17 +0100 |
Chris Cannam |
Implement and test sparse resizedTo; some tidying
|
| Wed, 22 May 2013 13:54:15 +0100 |
Chris Cannam |
Rename and reorder args of at and slice functions in matrix and vector for consistency with std module equivalents
|
| Wed, 22 May 2013 09:22:10 +0100 |
Chris Cannam |
Make concat also return sparse when all its inputs are sparse
|
| Wed, 22 May 2013 08:56:51 +0100 |
Chris Cannam |
Replace matrix thresholded with a more general filter function that works for sparse matrices too
|
| Tue, 21 May 2013 22:37:28 +0100 |
Chris Cannam |
Make sum, difference, scaled and abs use sparse operations and return sparse matrices if all inputs are sparse
|
| Tue, 21 May 2013 22:36:39 +0100 |
Chris Cannam |
Implement add, subtract, scaled, abs
|
| Tue, 21 May 2013 21:58:34 +0100 |
Chris Cannam |
Add general enumerate, ensure zeros fed in to makeSparse are not retained in non-zero lists; test
|
| Tue, 21 May 2013 17:40:33 +0100 |
Chris Cannam |
Using hashes is simpler, but turns out to be mostly no faster and sometimes much slower. Not one to merge back.
sparse
|
| Tue, 21 May 2013 16:00:21 +0100 |
Chris Cannam |
Merge from default branch
sparse
|
| Tue, 21 May 2013 14:29:22 +0100 |
Chris Cannam |
Another minor improvement
|
| Tue, 21 May 2013 12:11:40 +0100 |
Chris Cannam |
Proper sparse*sparse multiply, more efficient makeSparse. Still loads of room for improvement, but the timings are more predictable now at least.
|
| Mon, 20 May 2013 22:17:35 +0100 |
Chris Cannam |
Merge from sparse branch
|
| Mon, 20 May 2013 22:17:19 +0100 |
Chris Cannam |
A reasonable sparse multiply, and a bit quicker dense one
sparse
|
