Mercurial > hg > svcore

log

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
less more | rev 1305: (0) -1000 -300 -100 -60 +60 +100 +300 tip
age author description
2016-11-29 Chris Cannam Add gapless flag to MP3FileReader, and implement trimming the delay samples from the start (padding is not yet trimmed from end) mp3-gapless
2016-11-29 Chris Cannam Add test for last change to ColumnOp 3.0-integration
2016-11-28 Chris Cannam Further fix to decision about when not to interpolate in ColumnOp::distribute 3.0-integration
2016-11-28 Chris Cannam Merge 3.0-integration
2016-11-28 Chris Cannam Interpolate always, when interpolation requested. Previous test failed to take into account varying bin height in log-scale modes, which meant interpolation might be necessary even when more bins than pixels. 3.0-integration
2016-11-25 Chris Cannam Make CSV feature writer produce the same output with Qt 5.6 and 5.7+ 3.0-integration
2016-11-25 Chris Cannam Don't show this... We should find a better way to make it available in the UI 3.0-integration
2016-11-25 Chris Cannam Avoid being locale-specific in XSD parse 3.0-integration
2016-11-25 Chris Cannam Merge 3.0-integration
2016-11-25 Chris Cannam Improve mp3 offset detection, add test for truncated decode 3.0-integration
2016-11-25 Chris Cannam Don't always print these 3.0-integration
2016-11-25 Chris Cannam Merge from branch subdivide-instants 3.0-integration
2016-11-25 Chris Cannam Tidy up, with some convenient C++11isms subdivide-instants
2016-11-25 Chris Cannam Add Winnow as well subdivide-instants
2016-11-25 Chris Cannam Implement "Subdivide Selected Instants". Also add the relabel command to the history, which I previously forgot to do! subdivide-instants
2016-11-24 Chris Cannam Add MAD_BUFFER_GUARD padding at end of mp3 buffer, in order to ensure last frame is decoded successfully (otherwise the decoded audio is truncated). Another thing learned from madplay. 3.0-integration
2016-11-24 Chris Cannam Do that only for the first frame 3.0-integration
2016-11-24 Chris Cannam Filter out Xing/LAME info frames, rather than letting them go to the mp3 decoder as if they were audio frames. Fixes the 1152-sample zero pad at start of some decoded mp3 files (distinct from decoder delay). The logic here is based on the madplay code. 3.0-integration
2016-11-23 Chris Cannam Tag load diagnostic 3.0-integration
2016-11-23 Chris Cannam A bit more diagnostic info 3.0-integration
2016-11-23 Chris Cannam More diagnostic output 3.0-integration
2016-11-23 Chris Cannam More error reporting updates 3.0-integration
2016-11-23 Chris Cannam Warn when out-of-range bin requested 3.0-integration
2016-11-23 Chris Cannam Avoid potentially trying to allocate negative-sized vector! 3.0-integration
2016-11-23 Chris Cannam Debug note when we overrun end of data 3.0-integration
2016-11-23 Chris Cannam Trivial loop reordering for sequential index 3.0-integration
2016-11-22 Chris Cannam Debug output updates 3.0-integration
2016-11-22 Chris Cannam Avoid compiler warning 3.0-integration
2016-11-22 Chris Cannam Actually use the StorageAdviser output properly 3.0-integration
2016-11-22 Chris Cannam More informative debug output from StorageAdviser 3.0-integration
2016-11-22 Chris Cannam Add SVCERR and the ability to silence debug output (giving Sonic Annotator a --quiet mode) 3.0-integration
2016-11-21 Chris Cannam Hit counts optional 3.0-integration
2016-11-21 Chris Cannam Merge from default branch 3.0-integration
2016-11-21 Chris Cannam Send RDF-indexer warnings to debug rather than cerr
2016-11-21 Chris Cannam Merge from branch bqfft 3.0-integration
2016-11-21 Chris Cannam Branch to use bqfft code and remove FFTapi bqfft
2016-11-18 Chris Cannam Don't show this log by default 3.0-integration
2016-11-18 Chris Cannam Fix fairly obscure mis-canonicalisation in RealTime (leading to non-canonical but still correct values) 3.0-integration
2016-11-17 Chris Cannam Further column op tests and fixes 3.0-integration
2016-11-17 Chris Cannam Tests and fixes for distribute(). Although this version of interpolated distribution passes these tests, it isn't right visually -- the expected values in the tests are offset. To be continued. 3.0-integration
2016-11-17 Chris Cannam Start tests for ColumnOp (+ some resulting fixes) 3.0-integration
2016-11-16 Chris Cannam Debug output improvements, and make the checker actually attempt to call the descriptor function for known plugin types 3.0-integration
2016-11-15 Chris Cannam Test fixes 3.0-integration
2016-11-15 Chris Cannam RealTime fixes to comply with new, more demanding tests; also run same (failing) tests on Vamp SDK version 3.0-integration
2016-11-14 Chris Cannam Remove some no-longer-used project files 3.0-integration
2016-11-14 Chris Cannam Rearrange svcore tests so they can all be built from the top level 3.0-integration
2016-11-14 Chris Cannam Further RealTime test (currently failing) 3.0-integration
2016-11-10 Chris Cannam More timings and cache hit counts 3.0-integration
2016-11-10 Chris Cannam A note on cache hit findings 3.0-integration
2016-11-09 Chris Cannam Print out cache hit/miss counts 3.0-integration
2016-11-05 Chris Cannam Merge from branch piper 3.0-integration
2016-11-05 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
2016-11-04 Chris Cannam Fix omitted topmost bin 3.0-integration
2016-11-04 Chris Cannam Ensure columns all have correct height when retrieved (to avoid e.g. empty columns because a feature extractor's start time was >0) 3.0-integration
2016-11-04 Chris Cannam Merge from branch piper 3.0-integration
2016-11-04 Chris Cannam Some updates to accommodate adaptation for Sonic Annotator piper
2016-11-04 Chris Cannam Align Sonic Annotator with the new Piper-ified subrepos (bearing in mind we want neither Piper nor the plugin load checker in Sonic Annotator itself) piper
2016-11-03 Chris Cannam Update to use listargs variant of Piper stuff (so that the plugin winnowing feature from the penultimate commit actually works) piper
2016-11-03 Chris Cannam Make SVDEBUG always write to a log file -- formerly this was disabled in NDEBUG builds. I think there's little use to that, it just means that we keep adding more cerr debug output because we aren't getting the log we need. And SVDEBUG logging is not usually used in tight loops, I don't think the performance overhead is too serious. piper
2016-11-03 Chris Cannam Use plugin scan results to inform the list requests issued to Piper servers piper
less more | rev 1305: (0) -1000 -300 -100 -60 +60 +100 +300 tip