Commits · 3d25e71b407816098a421fe6b30edcd0f5af7ad1 · Alexander Traud / Opus · GitLab

Snippets Groups Projects

Oct 16, 2010
- More simplifications to the VQ search · 3d25e71b
  Jean-Marc Valin authored 14 years ago
  
  3d25e71b
- VQ search no longer needs to put more than one pulse at a time · aca7a630
  Jean-Marc Valin authored 14 years ago
  
  aca7a630
Oct 15, 2010
- tweaks to the split delta for transients · dfb3e686
  Jean-Marc Valin authored 14 years ago
  
  dfb3e686
- Transient detection fix for the case where a transient occurs during the overlap · 4d2d9fc9
  Jean-Marc Valin authored 14 years ago
  
  4d2d9fc9
- Tuning the bias for the L1 tf metric · 54fb7e56
  Jean-Marc Valin authored 14 years ago
  
  54fb7e56
- Unb0rked the VBR code: several units issues that made VBR completely useless · 8852563a
  Jean-Marc Valin authored 14 years ago
  
  8852563a
- Modifying the L1 metric for tf analysis by merging together all frequencies. · 8200b2d1
  Jean-Marc Valin authored 14 years ago
  
  Still need to make it use the right channel in stereo.
  8200b2d1
Oct 13, 2010
- Making it possible to have two transients in a row for longer frames · 4a8c1f11
  Jean-Marc Valin authored 14 years ago
  
  4a8c1f11
- Adding a high-pass filter in front of the transient analysis code · 2794b630
  Jean-Marc Valin authored 14 years ago
  
  2794b630
- No longer filling the stereo side with noise. · ccd51fe9
  Jean-Marc Valin authored 14 years ago
  
  ccd51fe9
- Taking into account TF decisions for VBR · ccd5a61a
  Jean-Marc Valin authored 14 years ago
  
  ccd5a61a
- New TF decision code based on L1-norm. Needs more work. · 0f8fc0be
  Jean-Marc Valin authored 14 years ago
  
  0f8fc0be
- Don't take into account silence for the coarse energy means · 844f00de
  Jean-Marc Valin authored 14 years ago
  
  844f00de
Oct 12, 2010
- Remove lower bound on bit-rate for tf signalling · a8160dde
  Jean-Marc Valin authored 14 years ago
  
  a8160dde
Oct 10, 2010
- Dynamic allocation before VBR · f1fea66b
  Jean-Marc Valin authored 14 years ago
  
  f1fea66b
- Retrained coarse energy mean and beta coefficients · 8226ac08
  Jean-Marc Valin authored 14 years ago
  
  8226ac08
Oct 04, 2010
- Making the allocation slightly less insane at very high rate · 8ccda887
  Jean-Marc Valin authored 14 years ago
  
  8ccda887
- Using high-band noise less aggressively · b441a157
  Jean-Marc Valin authored 14 years ago
  
  b441a157
Sep 30, 2010
- Allowing to change the allocation dynamically. · 6bf04627
  Jean-Marc Valin authored 14 years ago
  
  Uses a scaling factor that gets applied to the allocation matrix. Conflicts: libcelt/celt.c libcelt/rate.c libcelt/rate.h
  6bf04627
- Updates tandem-test for more robust error reporting and additional frame sizes and sample rates. · 4a11daa2
  Gregory Maxwell authored 14 years ago and Jean-Marc Valin committed 14 years ago
  
  4a11daa2
- Fixes a ibudget bust and a divide-by-zero at very low bit-rate · 83c26fbd
  Jean-Marc Valin authored 14 years ago
  
  83c26fbd
Sep 29, 2010
- Fixes a fixed-point overflow on 16-bit architectures · b1fe4f60
  Jean-Marc Valin authored 14 years ago
  
  b1fe4f60
- Adds a simple dynamic allocation scheme · d74c851f
  Jean-Marc Valin authored 14 years ago
  
  This should help on highly tonal signals by increasing the rate of bands with a lot more energy than the neighbouring bands.
  d74c851f
Sep 28, 2010
- Preventing bands from being coded at a rate below (for now) 3/8 bit/sample · b801da5e
  Jean-Marc Valin authored 14 years ago
  
  b801da5e
Sep 24, 2010
- Include tuning from exp_vorbis_tuning (disabled for now) · 78ea9fd0
  Jean-Marc Valin authored 14 years ago
  
  78ea9fd0
Sep 15, 2010
- Fixes a minor folding bug for N=2 bands in stereo · 8abc6bee
  Jean-Marc Valin authored 14 years ago
  
  8abc6bee
Sep 14, 2010
- Add an extra safety bit in intra_decision() · cb2c77c6
  Jean-Marc Valin authored 14 years ago
  
  cb2c77c6
- Fixes rare overflow in intra_decision() · ba760bc5
  Jean-Marc Valin authored 14 years ago
  
  ba760bc5
- Fixes fixed-point overflow in stereo_merge() · 3c15659c
  Jean-Marc Valin authored 14 years ago
  
  3c15659c
- Fixes a fixed-point overflow in haar1() · d60e0db0
  Jean-Marc Valin authored 14 years ago
  
  d60e0db0
- Speed up the bit allocator interpolation · 9ce95056
  Jean-Marc Valin authored 14 years ago
  
  9ce95056
- Fix stereo for N=2 · 5f3ad5f9
  Jean-Marc Valin authored 14 years ago
  
  5f3ad5f9
- Copy lowband when needed instead of undoing the transforms · 79b45338
  Jean-Marc Valin authored 14 years ago
  
  79b45338
Sep 13, 2010
- Allowing a max of 8 bits to qtheta · ad7d5e00
  Jean-Marc Valin authored 14 years ago
  
  ad7d5e00
- More quant_band() reorganisation. · a5ce03fe
  Jean-Marc Valin authored 14 years ago
  
  a5ce03fe
- Restructuring in quant_band() · a18c75e5
  Jean-Marc Valin authored 14 years ago
  
  a18c75e5
- Fixing tests on machines with no __builtin_clz() · 293b9f19
  Jean-Marc Valin authored 14 years ago
  
  293b9f19
- Separates stereo_band_mix() into the intensity and MS stereo cases · 35780760
  Jean-Marc Valin authored 14 years ago
  
  35780760
- Added stereo_merge(), which does the renormalisation too · 90e7fdd7
  Jean-Marc Valin authored 14 years ago
  
  90e7fdd7
- MacOS X typedef fix · 80cae178
  John Ridges authored 14 years ago
  
  80cae178