summaryrefslogtreecommitdiff
Commit message (Expand)AuthorAgeFilesLines
* Reverts the transient_analysis() changes from 4e8b9905exp_wip4Jean-Marc Valin2012-05-031-68/+40
* Fixes issues when downmixing to mono or when encoding below fullbandJean-Marc Valin2012-03-051-4/+5
* Tonality and pitch tuningJean-Marc Valin2012-03-053-12/+21
* Converting most of the new code to fixed-point (not complete yet)Jean-Marc Valin2012-03-052-63/+92
* Fixes some issues discovered with a log sweepJean-Marc Valin2012-03-052-24/+54
* Tuning the dynalloc adjustments and checking for LF transientsJean-Marc Valin2012-03-051-49/+42
* Adds a 3 Hz high-pass filter and boost allocation on leakageJean-Marc Valin2012-03-053-45/+50
* Dynalloc based on a bands that stand out of the "noise floor"Jean-Marc Valin2012-03-051-0/+70
* Transient/VBR tuning, give more bits to frames where pitch changesJean-Marc Valin2012-03-053-115/+110
* New transient code, weighted tonalityJean-Marc Valin2012-03-052-41/+109
* VBR tuningJean-Marc Valin2012-03-052-14/+19
* Adds automatic bandwidth detectionJean-Marc Valin2012-03-051-1/+32
* Better training setJean-Marc Valin2012-03-051-56/+56
* Adds low-energy feature, training with noiseJean-Marc Valin2012-03-053-141/+104
* Better trainingJean-Marc Valin2012-03-051-62/+120
* Adds MFCC standard deviation featuresJean-Marc Valin2012-03-053-51/+93
* Down to 17 features with same error rateJean-Marc Valin2012-03-052-72/+53
* Adds cepstral mean subtraction (CMS) to analysisJean-Marc Valin2012-03-052-66/+68
* Continuous trim calculations, tapset/spread now use new analysisJean-Marc Valin2012-03-053-15/+39
* Speech/music discrimination (not used for anything yet)Jean-Marc Valin2012-03-059-2/+946
* Cleanup, slightly changed analysis band layout, MFCC-like featuresJean-Marc Valin2012-03-052-10/+72
* Cleans up the most ugly parts of the analysis codeJean-Marc Valin2012-03-054-108/+166
* Adds an analysis function to control VBRJean-Marc Valin2012-03-054-54/+352
* Implements a better transient metric for VBRJean-Marc Valin2012-03-051-11/+21
* Using a "pure L1" metric for tf analysisJean-Marc Valin2012-03-051-20/+10
* Unconstrained VBR no longer attempt to hit an average rateJean-Marc Valin2012-03-051-3/+8
* Fix documenation typos.Ralph Giles2012-03-051-13/+13
* Fix typos in the API documentation.Ralph Giles2012-03-051-2/+2
* Restore the windows-side #ifdef check in silk/MacroDebug.h.Ralph Giles2012-03-051-1/+1
* Check for underflow in the debug USUB32.Ralph Giles2012-03-051-0/+4
* Enable the silk debug macros with FIXED_DEBUG.Ralph Giles2012-03-051-1/+1
* Fix a signed-compare warning.Ralph Giles2012-03-051-2/+2
* Print the actual arguments to the debug UADD/SUB32 routines.Ralph Giles2012-03-051-4/+4
* Uncomment the reporting for UADD32 and USUB32 in fixed_debug.h.Ralph Giles2012-03-051-2/+2
* Used unsigned intermediates with unsigned arguments in fixed_debug.h.Ralph Giles2012-03-051-2/+2
* Use automake's support for parallel test execution.Ralph Giles2012-03-051-1/+1
* Fix --enable-fixed-point-debug.Ralph Giles2012-03-054-12/+5
* Guard against pathologically small bufferVincent Penquerc'h2012-03-051-0/+5
* Make input mapping parameter const in multistream APIVincent Penquerc'h2012-03-052-8/+8
* Adds missing RESTORE_STACK statements on error pathsJean-Marc Valin2012-03-052-5/+40
* More consistent types for 16-bit architecturesJean-Marc Valin2012-03-053-5/+5
* Removes a number of macro definitions which are used nowhere in the codebase.Gregory Maxwell2012-03-059-31/+1
* Various multistream fixes.Gregory Maxwell2012-03-051-6/+15
* 40/60ms MDCT/Hybrid were not able to reach maximum bitrate. Now they can.Gregory Maxwell2012-03-054-48/+49
* Add config.h (and copyright notice) to repacketizer_demo.c.Gregory Maxwell2012-03-051-0/+29
* Eliminates an unused parameter warning in anti_collapse()Jean-Marc Valin2012-03-053-4/+4
* Last updates for draft -11v0.9.9Jean-Marc Valin2012-02-1712-812/+1176
* Modifies the comparison tool to make it much more permissive.Jean-Marc Valin2012-02-091-129/+165
* Fixes MSVC warningsKoen Vos2012-01-311-3/+2
* Tweaks the CELT fractional resampling delay to get perfect alignmentJean-Marc Valin2012-01-314-13/+10