Fri, 16 Aug 2019 12:28:58 +0100 |
Chris Cannam |
Some experimental feature adjustments - try using diffs between chroma cells (no better at first glance)
feature-alternatives tip
|
Thu, 11 Jul 2019 16:29:40 +0100 |
Chris Cannam |
Add AppVeyor link
default
|
Thu, 11 Jul 2019 16:12:07 +0100 |
Chris Cannam |
Vamp SDK sources
|
Thu, 11 Jul 2019 15:04:13 +0100 |
Chris Cannam |
Tidy MS build bits
|
Thu, 11 Jul 2019 15:03:59 +0100 |
Chris Cannam |
Plural
|
Thu, 11 Jul 2019 14:54:29 +0100 |
Chris Cannam |
Merge
|
Thu, 11 Jul 2019 14:54:13 +0100 |
Chris Cannam |
Incomplete MSVC build
|
Thu, 11 Jul 2019 11:52:54 +0100 |
Chris Cannam |
vorbis-tools, for oggdec
|
Thu, 11 Jul 2019 11:17:34 +0100 |
Chris Cannam |
Further README tweak and build fix
|
Thu, 11 Jul 2019 10:58:18 +0100 |
Chris Cannam |
Further update to README
|
Thu, 11 Jul 2019 10:52:15 +0100 |
Chris Cannam |
Obsolete test files
|
Thu, 11 Jul 2019 10:51:58 +0100 |
Chris Cannam |
Update README
|
Thu, 11 Jul 2019 10:41:11 +0100 |
Chris Cannam |
MLton needed for vamp-plugin-tester build
|
Thu, 11 Jul 2019 09:59:26 +0100 |
Chris Cannam |
Remove bulk RDF
|
Thu, 11 Jul 2019 09:58:56 +0100 |
Chris Cannam |
Revert to having only a single plugin - I think this should be sufficiently backward compatible after all
|
Thu, 11 Jul 2019 09:19:21 +0100 |
Chris Cannam |
Minor tweaks
|
Thu, 11 Jul 2019 08:57:51 +0100 |
Chris Cannam |
Ah, I see - it's the install target that is depending on programs. Oh well, let's just add this library after all
|
Wed, 10 Jul 2019 19:13:56 +0100 |
Chris Cannam |
I think we can allow this many
|
Wed, 10 Jul 2019 19:10:25 +0100 |
Chris Cannam |
Not sure why that option wasn't working?
|
Wed, 10 Jul 2019 18:52:36 +0100 |
Chris Cannam |
Make BulkTuningDifference much faster when working with more than two files in fine-tuning mode, by storing the reference signal and calculating the fine-tuned features from that (shifted the opposite way) while leaving the other features alone at their coarse rotations. Slightly changes the results but not by a significant amount, I think.
|
Wed, 10 Jul 2019 14:00:51 +0100 |
Chris Cannam |
Add category
|
Wed, 10 Jul 2019 12:00:28 +0100 |
Chris Cannam |
Minor fixes, expand tests, and update RDF description
|
Tue, 09 Jul 2019 16:14:19 +0100 |
Chris Cannam |
Add BulkTuningDifference plugin
|
Tue, 09 Jul 2019 15:09:33 +0100 |
Chris Cannam |
Switch from 60bpo to 120 - seems to work better (?) with little performance downside. Also avoid recalculating second feature if not needed for fine tuning, and extend regression test
|
Tue, 09 Jul 2019 12:34:25 +0100 |
Chris Cannam |
Very simplistic regression test
|
Thu, 13 Jun 2019 11:32:01 +0100 |
Chris Cannam |
Make fine tuning optional; wire up forgotten max range option!
|
Thu, 16 May 2019 15:56:07 +0100 |
Chris Cannam |
Couple more parameters
|
Wed, 15 May 2019 17:52:47 +0100 |
Chris Cannam |
Fix a couple of compiler warnings
|
Mon, 17 Jul 2017 15:46:04 +0100 |
Chris Cannam |
Metadata
|
Mon, 17 Jul 2017 15:35:03 +0100 |
Chris Cannam |
Add necessary header
|
Thu, 19 May 2016 13:54:23 +0100 |
Chris Cannam |
Add constant-q-cpp code
|
Thu, 19 May 2016 13:52:48 +0100 |
Chris Cannam |
No, git modules are just too horrible. Let's pull in the code instead. Sigh
|
Thu, 19 May 2016 13:31:44 +0100 |
Chris Cannam |
Experiment with git submodules (as well as hg subrepos!)
|
Thu, 19 May 2016 13:31:10 +0100 |
Chris Cannam |
Comment only
|
Fri, 04 Dec 2015 11:07:31 +0000 |
Chris Cannam |
Reduce max frequency difference to 4 semitones. This is because we can very easily get errors of 5 semitones (a 5 semitone error "down" being equivalent to a perfect fifth "up").
|
Fri, 27 Nov 2015 16:58:24 +0000 |
Chris Cannam |
Fix error from tester
|
Fri, 25 Sep 2015 09:54:00 +0100 |
Chris Cannam |
Update subrepo
|
Thu, 10 Sep 2015 15:58:11 +0100 |
Chris Cannam |
Update subrepo
|
Fri, 17 Jul 2015 15:50:02 +0100 |
Chris Cannam |
Subrepo update
|
Fri, 17 Jul 2015 15:48:15 +0100 |
Chris Cannam |
Update subrepo
|
Fri, 27 Mar 2015 17:20:40 +0000 |
Chris Cannam |
Add max-duration parameter. The results are often going to be quite wrong if you use this, simply because the two performances may be at different speeds and so the same duration of each may represent a different (sub-)performance. But it's sometimes necessary just to avoid blowing the thing up.
|
Fri, 13 Mar 2015 15:03:54 +0000 |
Chris Cannam |
Comment
|
Mon, 16 Feb 2015 15:38:07 +0000 |
Chris Cannam |
Use library path
|
Wed, 11 Feb 2015 14:23:03 +0000 |
Chris Cannam |
Minor note
|
Fri, 06 Feb 2015 09:03:12 +0000 |
Chris Cannam |
Remove some (but not yet all) debug output
|
Fri, 06 Feb 2015 09:01:34 +0000 |
Chris Cannam |
Some speed improvements
|
Thu, 05 Feb 2015 10:38:19 +0000 |
Chris Cannam |
Example
|
Thu, 05 Feb 2015 10:33:14 +0000 |
Chris Cannam |
Minor wording change
|
Thu, 05 Feb 2015 10:13:31 +0000 |
Chris Cannam |
Rearrange code, include subrepo etc
|
Thu, 05 Feb 2015 09:52:28 +0000 |
Chris Cannam |
Update to reflect changes in Chromagram api
|
Thu, 05 Feb 2015 09:43:31 +0000 |
Chris Cannam |
Return values through plugin outputs
|
Thu, 05 Feb 2015 08:56:42 +0000 |
Chris Cannam |
Fix uninitialised values
|
Thu, 05 Feb 2015 08:55:47 +0000 |
Chris Cannam |
Fix typo
|
Thu, 05 Feb 2015 08:54:52 +0000 |
Chris Cannam |
Fine frequency search
|
Thu, 05 Feb 2015 08:38:59 +0000 |
Chris Cannam |
Use Chromagram from libcq, fix incorrect rotation
|
Wed, 04 Feb 2015 15:11:42 +0000 |
Chris Cannam |
Merge
|
Wed, 04 Feb 2015 15:10:07 +0000 |
Chris Cannam |
The bulk of the chroma-compare implementation. Should be using the newly refactored Chromagram class from libcq though (chroma is currently upside-down!)
|
Wed, 04 Feb 2015 10:15:57 +0000 |
Chris Cannam |
Rename & duplicate plugin
|
Tue, 03 Feb 2015 18:46:54 +0000 |
Chris Cannam |
Merge
|
Mon, 02 Feb 2015 18:56:41 +0000 |
Chris Cannam |
Merge
|