log transform/FeatureExtractionModelTransformer.cpp @ 1881:b504df98c3be

age author description
Fri, 26 Jun 2020 11:45:39 +0100 Chris Cannam Ensure completion on output model is started at zero, so if it's checked before the input model has become ready and the transform has begun, it is not accidentally reported as complete (affected re-aligning models in Sonic Lineup when replacing the session)
Thu, 09 Apr 2020 11:22:55 +0100 Chris Cannam Unify various vectors to our base floatvec_t type; store columns in fft model cache at their desired height so we can return a reference (speeding up the peak-frequency spectrogram in particular)
Fri, 20 Mar 2020 16:30:33 +0000 Chris Cannam Use shared_ptr for plugin instances throughout audio-source-refactor
Tue, 17 Sep 2019 11:21:33 +0100 Chris Cannam More graceful handling of failure to construct FFT models in the case where the source model has already been deleted before this occurs
Tue, 10 Sep 2019 16:34:47 +0100 Chris Cannam Split EditableDenseThreeDimensionalModel into explicitly compressed and uncompressed variants. Simplifies the uncompressed version, and we may want to consider whether we need the compressed one at all.
Wed, 04 Sep 2019 17:31:02 +0100 Chris Cannam Some debug tweaks
Thu, 18 Jul 2019 14:42:24 +0100 Chris Cannam Respect use-flexi-note-model setting
Wed, 17 Jul 2019 11:19:46 +0100 Chris Cannam Conditionally compile debug out by-id
Mon, 15 Jul 2019 14:12:37 +0100 Chris Cannam Deleted-model checks, & debug output by-id
Fri, 05 Jul 2019 16:55:54 +0100 Chris Cannam Don't hold on to borrowed pointer around the loop - so as to be informed when it becomes obsolete by-id
Fri, 05 Jul 2019 15:28:07 +0100 Chris Cannam Further review of ById: make IDs only available when adding a model to the ById store, not by querying the item directly. This means any id encountered in the wild must have been added to the store at some point (even if later released), which simplifies reasoning about lifecycles by-id
Tue, 02 Jul 2019 11:49:28 +0100 Chris Cannam Update FFT model to ById by-id
Wed, 26 Jun 2019 17:25:20 +0100 Chris Cannam More work on transformers by-id
Wed, 26 Jun 2019 14:59:09 +0100 Chris Cannam Some work on models and transformers by-id
Wed, 20 Mar 2019 16:22:13 +0000 Chris Cannam Start updating SparseOneDimensionalModel single-point
Mon, 18 Mar 2019 14:17:20 +0000 Chris Cannam Introduce deferred notifier, + start converting sparse time-value model (perhaps we should rename it too) single-point
Fri, 15 Mar 2019 14:23:50 +0000 Chris Cannam Update RegionModel following NoteModel, er, model. They have quite a bit in common that we should now pull out some of single-point
Thu, 14 Mar 2019 15:31:59 +0000 Chris Cannam Rework NoteModel commands (not entirely successfully); remove FlexiNoteModel as it has always been almost entirely identical to NoteModel (unlike its layer counterpart) single-point
Wed, 13 Mar 2019 16:00:13 +0000 Chris Cannam Further API updates single-point
Wed, 13 Mar 2019 14:50:10 +0000 Chris Cannam Update NoteModel to use EventSeries. This is incomplete and won't provide enough to update svgui for it yet; must also refactor to avoid duplication of nasty APIs when updating the rest of the models single-point
Mon, 26 Nov 2018 14:33:41 +0000 Chris Cannam Use nullptr throughout fix-static-analysis
Thu, 18 Oct 2018 13:14:56 +0100 Chris Cannam Some debug + tidying
Thu, 01 Mar 2018 18:02:22 +0000 Chris Cannam Untabify
Tue, 07 Feb 2017 14:55:19 +0000 Chris Cannam Some work to make spectrum layers (and slice layers generally) zoomable in the frequency axis. Also fixes a number of view id mixups in SliceLayer which broke offset calculations for the x axis scale.
Mon, 06 Feb 2017 09:18:44 +0000 Chris Cannam Catch exception from initialise
Tue, 31 Jan 2017 11:13:20 +0000 Chris Cannam Handle exceptions raised during plugin processing (as opposed to on init/configure, which were already handled)
Thu, 19 Jan 2017 11:01:41 +0000 Chris Cannam Fix absurd typo that was preventing transforms from running at all!
Wed, 18 Jan 2017 14:21:39 +0000 Chris Cannam Set abandoned flag when bailing out of run() because init failed at the start; then check that flag in awaitOutputModels() so as not to get stuck on a wait condition
Wed, 16 Nov 2016 16:12:42 +0000 Chris Cannam Debug output improvements, and make the checker actually attempt to call the descriptor function for known plugin types 3.0-integration
Sat, 05 Nov 2016 10:40:22 +0000 Chris Cannam Avoid assigning model resolution of 0 to models coming from plugins that report weirdly high output rates (symptom: data did not show up in edit layer data dialog, as the row/frame conversion failed) piper
Fri, 28 Oct 2016 11:32:36 +0100 Chris Cannam Ensure plugin is deleted in same thread that created and used it project-file-rework
Fri, 21 Oct 2016 14:51:22 +0100 Chris Cannam Fixes to choice of plugin factory piper
Fri, 21 Oct 2016 11:49:27 +0100 Chris Cannam Restore native-Vamp factory and make the choice between Piper and Native a preference piper
Mon, 17 Oct 2016 14:18:23 +0100 Chris Cannam Pull out model creation into the transformer thread run(), so that all communications with the plugin server happen on a single thread. Then make the model accessor wait for them to be created (which still happens right at the start of processing) before returning. piper
Fri, 22 Jan 2016 13:39:45 +0000 Chris Cannam Convert 3d model column type from QVector to std::vector; replace another user of ResizeableBitset 3.0-integration
Mon, 15 Jun 2015 12:19:47 +0100 Chris Cannam Rework audio file reader API to prefer using std containers simple-fft-model
Fri, 12 Jun 2015 14:51:46 +0100 Chris Cannam Gut the old code, but don't replace it yet (so nothing will link yet) simple-fft-model
Wed, 10 Jun 2015 17:06:02 +0100 Chris Cannam Tidy dense time-value model API a bit; add first simple unit test for FFT model
Wed, 10 Jun 2015 13:12:06 +0100 Chris Cannam Tighten up error handling
Thu, 14 May 2015 14:04:41 +0100 Chris Cannam Handle case where plugin erroneously returns sample rate 0 for fixed-rate output
Wed, 04 Mar 2015 12:01:04 +0000 Chris Cannam Introduce and use a samplerate type cxx11
Tue, 03 Mar 2015 17:09:19 +0000 Chris Cannam This code now compiles. Main problem: sample rate types cxx11
Fri, 14 Nov 2014 13:51:33 +0000 Chris Cannam When using an aggregate model to pass data to a transform, zero-pad the shorter input to the duration of the longer rather than truncating the longer. (This is better behaviour for e.g. MATCH, and in any case the code was previously truncating incorrectly and ending up with garbage data at the end.)
Wed, 03 Sep 2014 11:22:46 +0100 Chris Cannam Fix memory leak
Wed, 18 Jun 2014 08:34:46 +0100 Chris Cannam Merge from branch tony_integration warnfix_no_size_t
Tue, 17 Jun 2014 14:49:49 +0100 Chris Cannam More size_t evisceration and warning fixes warnfix_no_size_t
Fri, 13 Jun 2014 17:39:01 +0100 Chris Cannam Add an experimental cancel button to view progress bars tonioni
Wed, 02 Apr 2014 08:51:12 +0100 Chris Cannam Merge from branch tonioni tony_integration
Tue, 01 Apr 2014 13:27:32 +0100 Chris Cannam Comment out some debug tonioni
Tue, 11 Mar 2014 17:32:31 +0000 Chris Cannam Merge from default branch tony_integration
Tue, 11 Mar 2014 17:30:35 +0000 Chris Cannam Merge from branch tonioni tony_integration
Tue, 11 Mar 2014 17:29:44 +0000 Chris Cannam Use fixed-rate feature no directly as column number, rather than do a calculation that might introduce rounding error
Wed, 29 Jan 2014 15:58:15 +0000 Chris Cannam Callback handler for additional models at end of processing tonioni
Wed, 29 Jan 2014 09:31:22 +0000 Chris Cannam For outputs with unknown bin count or multiple bins with variable sample rate, create additional output models for bins beyond the first tonioni
Tue, 28 Jan 2014 18:52:22 +0000 Chris Cannam Use a sparse time-value model only for outputs with fixed bin count of 1, not for those with unknown bin count. tonioni
Tue, 28 Jan 2014 15:01:54 +0000 Chris Cannam Merge from tonioni branch tony_integration
Wed, 11 Dec 2013 21:40:27 +0000 Chris Cannam Fix handling of negative frame times in returned features from plugin -- we don't handle them, so don't insert them into the model (instead of blowing up)
Wed, 04 Dec 2013 19:30:02 +0000 Chris Cannam Use a settings setting for flexi/non-flexi determination -- this may not be the right thing in the long run but it's simpler and easier than passing through a random value that doesn't actually come from anywhere tonioni
Mon, 02 Dec 2013 12:29:09 +0000 Chris Cannam Update the rest of the transforms stuff, so now it builds and runs again tonioni_multi_transform
Mon, 02 Dec 2013 11:17:24 +0000 Chris Cannam More on multi-transform stuff tonioni_multi_transform
Tue, 26 Nov 2013 14:37:01 +0000 Chris Cannam Merge from default branch tonioni
Tue, 26 Nov 2013 13:35:08 +0000 Chris Cannam Drop std:: from cout, cerr, endl -- pull these in through Debug.h
Fri, 12 Jul 2013 13:10:28 +0100 Chris Cannam Merge from the default branch tonioni
Thu, 09 May 2013 17:31:31 +0100 Chris Cannam Merge qt5
Thu, 09 May 2013 17:30:44 +0100 Chris Cannam Remove some debug qt5
Tue, 16 Apr 2013 15:45:25 +0100 Chris Cannam Remove debug
Tue, 16 Apr 2013 15:20:57 +0100 Chris Cannam Remove some debug
Sat, 13 Apr 2013 15:36:54 +0100 gyorgyf cleanup tonioni
Sat, 13 Apr 2013 13:39:17 +0100 gyorgyf fixed flexi note layer and propagated model selection to analyser tonioni
Thu, 28 Mar 2013 17:08:04 +0000 Chris Cannam Warn if the output sample rate is higher than that of the input model
Wed, 27 Mar 2013 16:53:31 +0000 Chris Cannam Remove no-longer-used code
Wed, 27 Mar 2013 15:27:33 +0000 Chris Cannam Fix calculation of timestamps for fixed-rate features
Wed, 27 Mar 2013 15:07:27 +0000 Chris Cannam outputFeatureNo -> outputNo
Wed, 27 Mar 2013 14:51:49 +0000 Chris Cannam Fix incorrect retrieval of point in model by row number if point frame values fell between model resolution boundaries
Mon, 17 Oct 2011 11:57:25 +0100 Chris Cannam Minor debug change
Tue, 14 Jun 2011 15:26:52 +0100 Chris Cannam Seems to be a bad idea to use plain DEBUG symbol on OS/X (system wants it)
Mon, 16 May 2011 17:19:09 +0100 Chris Cannam Convert many cerrs to DEBUGs debug-output
Thu, 12 May 2011 17:31:24 +0100 Chris Cannam Remove most toStdString calls (no longer needed, with debug header) debug-output
Wed, 04 May 2011 14:05:08 +0100 Chris Cannam * Introduce ResourceFinder
Thu, 14 Apr 2011 15:20:27 +0100 Chris Cannam Make more (though still not entirely) robust in the face of running out of disc space
Mon, 08 Jun 2009 15:40:01 +0000 Chris Cannam * Use notes layer for things with MIDI units as well as Hz
Thu, 12 Feb 2009 15:26:43 +0000 Chris Cannam * Some fairly simplistic code to set up layer type properties based on
Tue, 10 Feb 2009 17:24:30 +0000 Chris Cannam * Fix crash when processing a file that is shorter than a single block
Mon, 09 Feb 2009 11:53:29 +0000 Chris Cannam * get whole columns at a time from fft model when running transform
Fri, 23 Jan 2009 14:00:29 +0000 Chris Cannam * make compression optional in editable dense 3d model, and some tweaks
Thu, 22 Jan 2009 17:58:33 +0000 Chris Cannam * try some qvectors
Fri, 28 Nov 2008 13:36:13 +0000 Chris Cannam Various fixes:
Fri, 07 Nov 2008 17:31:06 +0000 Chris Cannam * Update to new-line vamp plugin sdk
Wed, 29 Oct 2008 12:37:32 +0000 Chris Cannam * Ensure that region features are added even if they lack any values!
Thu, 09 Oct 2008 13:13:33 +0000 Chris Cannam * Support transforms that output regions with more than one bin (by
Wed, 08 Oct 2008 15:27:31 +0000 Chris Cannam * Fix (kinda) display of fixed-sample-rate features with sample rate
Thu, 25 Sep 2008 12:44:23 +0000 Chris Cannam * some work on transform finder
Thu, 18 Sep 2008 16:08:14 +0000 Chris Cannam * Add region model and layer; improve assignment of model types to
Thu, 10 Jul 2008 12:50:56 +0000 Chris Cannam * Updates for 1.3 sv1-v1.3 sv1-v1.3rc1
Wed, 18 Jun 2008 13:40:10 +0000 Chris Cannam * Fix crash when running Generator
Wed, 11 Jun 2008 12:53:20 +0000 Chris Cannam * Wait for input to be ready before running real-time effect -- even if
Thu, 08 May 2008 14:46:22 +0000 Chris Cannam * Speed up spectrogram painting by releasing mutex in FFTDataServer
Wed, 12 Mar 2008 18:02:17 +0000 Chris Cannam * Move plugin/transform to plain transform. This way transform can depend on base plugin/transform/FeatureExtractionModelTransformer.cpp@de1282742c53