- 98ec800 libOpus: fix integer overflow in silk_resampler_down2_hp by Neelkamal Semwal · 3 years, 3 months ago
- 09f8f05 Build time improvement, for MSVC use intrin0.h instead of intrin.h and remove usage of stdio.h in production code by Marcus Asteborg · 4 years, 4 months ago
- 57092f9 Improve silence handling by Jean-Marc Valin · 6 years ago
- a7aa7fb Fixing initialization issues on small frame sizes by Jean-Marc Valin · 6 years ago
- 843121b Fixes analysis buffering for silence and complexity changes by Jean-Marc Valin · 6 years ago
- 0619d08 Avoiding problems with x87 by Jean-Marc Valin · 6 years ago
- 5d29023 Retrain RNN classifier weights to include reverberated speech by Jean-Marc Valin · 6 years ago
- c1c247d Make bandwidth detection more conservative to avoid low-passing frames by Jean-Marc Valin · 6 years ago
- 8299edf Scaling back the pitch filter when most of the energy is above 3.2 kHz by Jean-Marc Valin · 7 years ago
- eee6898 fix float constants by Jean-Marc Valin · 7 years ago
- ff98202 Retrained speech/music RNN by Jean-Marc Valin · 7 years ago
- b30f45b Fixing (hopefully) bandwidth detection for 24 kHz analysis by Jean-Marc Valin · 7 years ago
- af93fbd Add RNN for VAD and speech/music classification by Jean-Marc Valin · 7 years ago
- dbe22f1 Working around misdetected audio bandwidth by Jean-Marc Valin · 7 years ago
- 5bb3cbc Add "f" suffix to float constants by Jean-Marc Valin · 7 years ago
- 4144342 Improving tonality estimation with delayed decision by Jean-Marc Valin · 7 years ago
- 1d7dea1 Initialize speech/music prob based on application by Jean-Marc Valin · 7 years ago
- 15a30e2 Properly init speech/music confidence values by Jean-Marc Valin · 7 years ago
- cf86d25 fix comments by Jean-Marc Valin · 7 years ago
- 7be054b Avoid using log2() since it's not in C90 by Jean-Marc Valin · 7 years ago
- 0cc4d96 Adding leakage modelling to boost bands by Jean-Marc Valin · 7 years ago
- a1c2d71 Analysis refactoring for fixed-point scaling by Jean-Marc Valin · 7 years ago
- d6d7037 Fix compiler warnings by Mark Harris · 8 years ago
- cf9409f Makes analysis run at 24 kHz, with 20-ms frames by Jean-Marc Valin · 8 years ago
- dbff5fc Prevent tonality estimator from running too often with 2.5 and 5 ms frames by Jean-Marc Valin · 8 years ago
- 2ff6556 Making stereo_itheta() use the same atan2() approximation as tonality_analysis() by Jean-Marc Valin · 8 years ago
- 3648134 New DTX that works in all modes (SILK/CELT/HYBRID) by Felicia Lim · 8 years ago
- fe4d91c Simplifying fast_atan2f() by Jean-Marc Valin · 8 years ago
- d68e809 Quality: Retrained classifier MLP with better data by Jean-Marc Valin · 9 years ago
- 9b1a27a Address static analysis reports from scan-build by Jean-Marc Valin · 8 years ago
- 9e3872a Add tonality_analysis_reset(). by Ralph Giles · 9 years ago
- d43445f Add an initializer for TonalityAnalysisState. by Ralph Giles · 9 years ago
- f48abe8 armv7(float): Optimize encode usecase using NE10 library by Viswanath Puttagunta · 9 years ago
- 8adff0b Fix const struct to work with C++ compilation by Mark Harris · 9 years ago
- 122971b More NaN hardening in the analysis code by Jean-Marc Valin · 11 years ago
- a71c9ad Fixes MSVC conversion warnings by Jean-Marc Valin · 11 years ago
- ca6fac0 Fixes some minor issues found by scan build by Jean-Marc Valin · 11 years ago
- 7830cf1 Replace "inline" with OPUS_INLINE. by Gregory Maxwell · 11 years ago
- 2925444 Makes analysis work when encoding more than 20 ms at a time by Jean-Marc Valin · 11 years ago
- a4c2512 run_analysis() doesn't need to return the frame size anymore by Jean-Marc Valin · 11 years ago
- b90e63b Moves frame size selection back out of opus_encode_native() by Jean-Marc Valin · 11 years ago
- ba7dbb3 Analysis scaling fixes by Jean-Marc Valin · 11 years ago
- fc1b1f9 Makes speech/music detection work with FIXED_POINT (code still float) by Jean-Marc Valin · 11 years ago
- 3ab03e0 First pass at making the analysis code run with FIXED_POINT by Jean-Marc Valin · 11 years ago
- 91904a4 Moves the analysis back to opus_encode_native() to it runs for all streams. by Jean-Marc Valin · 11 years ago
- 5280c71 Revert "Fixes a number of double promotions and missing casts." by Gregory Maxwell · 11 years ago
- 4327972 Fixes a number of double promotions and missing casts. by Gregory Maxwell · 11 years ago
- 4eb399d More speech/music detection comments (delayed decision) by Jean-Marc Valin · 11 years ago
- 0f5ff80 Commenting the speech/music Markov code by Jean-Marc Valin · 11 years ago
- e83a365 Fixes more MSVC warnings by Jean-Marc Valin · 11 years ago
- 16ba19a Fixes MSVC warnings by Jean-Marc Valin · 11 years ago
- 278389d Automatic bandwidth decisions get more conservative as rate increases. by Jean-Marc Valin · 11 years ago
- 6862b44 Slight simplification of the bandwidth detection code by Jean-Marc Valin · 11 years ago
- a8783a1 Converts the analysis to ALLOC() by Jean-Marc Valin · 11 years ago
- c37834b Fixes a minor bug found by valgrind in the bandwidth detector by Jean-Marc Valin · 11 years ago
- c5e04e4 Makes bandwidth detection less aggressive. by Jean-Marc Valin · 12 years ago
- 097fd4b Fixes C90 issues for obsolete compilers by Robert Meakins · 12 years ago
- 73142b1 Makes the speech/music probability estimation mode conservative by Jean-Marc Valin · 12 years ago
- 742aac1 Adds silence probability to speech/music detector by Jean-Marc Valin · 12 years ago
- 51f4a32 Adds support for delayed decision by Jean-Marc Valin · 12 years ago
- 48ac122 Makes analysis usable for all frame sizes by Jean-Marc Valin · 12 years ago
- 7ebacf4 Moves analysis to the beginning of opus_encode() by Jean-Marc Valin · 12 years ago
- d683c76 Fixes MSVC warnings for double->float and float->int conversion by Jean-Marc Valin · 12 years ago
- 7509fdb New bandwidth detection code by Jean-Marc Valin · 12 years ago
- 3b60e81 Fix the MSVC build. by Joshua Bowman · 12 years ago
- b2e19ec copysignf() was C99-only by Jean-Marc Valin · 12 years ago
- 0e49059 Gets rid of unused boost analysis code that never worked anyway. by Jean-Marc Valin · 12 years ago
- 317ffc2 Tuning the speech/music probability model by Jean-Marc Valin · 12 years ago
- 027ff07 Shrinks tansig table, makes data static const by Jean-Marc Valin · 12 years ago
- c5880fe Builds the analysis files more cleanly than #including C files by Jean-Marc Valin · 12 years ago
- b15b30c Speeds up analysis by getting rid of calls to atan2() and cos() by Jean-Marc Valin · 12 years ago
- 0892c16 Tonality and pitch tuning by Jean-Marc Valin · 13 years ago
- 2a9fdbc Transient/VBR tuning, give more bits to frames where pitch changes by Jean-Marc Valin · 13 years ago
- 971b055 New transient code, weighted tonality by Jean-Marc Valin · 13 years ago
- 70d90d1 VBR tuning by Jean-Marc Valin · 13 years ago
- a205457 Adds automatic bandwidth detection by Jean-Marc Valin · 13 years ago
- 7609b67 Adds low-energy feature, training with noise by Jean-Marc Valin · 13 years ago
- 747c817 Adds MFCC standard deviation features by Jean-Marc Valin · 13 years ago
- cd213ea Down to 17 features with same error rate by Jean-Marc Valin · 13 years ago
- ac3ac4a Adds cepstral mean subtraction (CMS) to analysis by Jean-Marc Valin · 13 years ago
- 9987a3b Speech/music discrimination (not used for anything yet) by Jean-Marc Valin · 13 years ago
- 73eb363 Cleanup, slightly changed analysis band layout, MFCC-like features by Jean-Marc Valin · 13 years ago
- e9c353a Cleans up the most ugly parts of the analysis code by Jean-Marc Valin · 13 years ago
- 1a2e765 Adds an analysis function to control VBR by Jean-Marc Valin · 13 years ago