Commits · d4018c3a22d3eee203b018ba751ae00fd3549182 · Alexander Traud / Opus · GitLab

Snippets Groups Projects

Feb 26, 2008
- Reduced code duplication in vq.c by adding mix_pitch_and_residual() which · d4018c3a
  Jean-Marc Valin authored 17 years ago
  
  is used by both by the encoding and decoding functions.
  d4018c3a
- C90-fying ectest · 2f5ccf60
  Jean-Marc Valin authored 17 years ago
  
  2f5ccf60
- Another 16-bit fix. · 7a455d4e
  Jean-Marc Valin authored 17 years ago
  
  7a455d4e
- fixed-point: got stereo to work again by fixing renormalise_bands() · b30dd9e1
  Jean-Marc Valin authored 17 years ago
  
  b30dd9e1
- fixed-point: celt_pgain_t now a 16-bit value (Q15) · 2805a14e
  Jean-Marc Valin authored 17 years ago
  
  2805a14e
- celt_mask_t for masking curves · 5f09ea56
  Jean-Marc Valin authored 17 years ago
  
  5f09ea56
- Removed useless masking curve for current frame and ignored Fs argument to · 8835a4df
  Jean-Marc Valin authored 17 years ago
  
  masking computation functions.
  8835a4df
- pitch gain is now celt_pgain_t · 017d4455
  Jean-Marc Valin authored 17 years ago
  
  017d4455
- fixed-point: band energy now a 32-bit value. It might have (barely) fix into · b60340f7
  Jean-Marc Valin authored 17 years ago
  
  16-bit, but at this point, it's not worth the trouble and loss of accuracy.
  b60340f7
- fixed-point: added a celt_ener_t type for band energy. · e901fe35
  Jean-Marc Valin authored 17 years ago
  
  e901fe35
- Minus a bunch of warnings when enabling alloca() · e046c18a
  Jean-Marc Valin authored 17 years ago
  
  e046c18a
- Put the channel count outside of the energy calculation. Increased the allowed · e995c31f
  Jean-Marc Valin authored 17 years ago
  
  range for the mono encoder.
  e995c31f
- fixed-point: celt_norm_t now a 16-bit value. · 2fa8affd
  Jean-Marc Valin authored 17 years ago
  
  2fa8affd
Feb 25, 2008
- 16-bit fixes and warning fixes · 9a5f3d2e
  Jean-Marc Valin authored 17 years ago
  
  9a5f3d2e
- Trying to be nice with 16-bit chips. · 508de38d
  Jean-Marc Valin authored 17 years ago
  
  508de38d
- Corrected the last digit of pi after massive public outcry :-) · 65d79e44
  Jean-Marc Valin authored 17 years ago
  
  65d79e44
- Ogg decoder does gapless for the beginning of the file. · b5a4c40d
  Jean-Marc Valin authored 17 years ago
  
  b5a4c40d
- fixed-point: Using a NORM_SCALING of 16384, sig_norm_t is still a float though. · b4dfce46
  Jean-Marc Valin authored 17 years ago
  
  b4dfce46
- fixed-point: celt_sig_t now a 32-bit value. · 5c032a4c
  Jean-Marc Valin authored 17 years ago
  
  5c032a4c
- fixed-point: initial support for using the fixed-point MDCT (rest is still all · 49ca99ef
  Jean-Marc Valin authored 17 years ago
  
  float)
  49ca99ef
- MDCT is in fixed-point now · bd43729e
  Jean-Marc Valin authored 17 years ago
  
  bd43729e
- MDCT conversion, part I. · 2e8a3b20
  Jean-Marc Valin authored 17 years ago
  
  2e8a3b20
Feb 24, 2008
- real fft no longer needs an internal buffer. · 42667b0a
  Jean-Marc Valin authored 17 years ago
  
  42667b0a
- real fft snr comparison testcase includes DC · 4508ebd5
  Jean-Marc Valin authored 17 years ago
  
  4508ebd5
- Float FFT now does the same scaling as the fixed-point FFT · 44830b04
  Jean-Marc Valin authored 17 years ago
  
  44830b04
- minor tweak to FFT · e8b6830f
  Jean-Marc Valin authored 17 years ago
  
  e8b6830f
- Added a mixed-precision version of the FFT with 32-bit data and 16-bit twiddles. · d911bc4d
  Jean-Marc Valin authored 17 years ago
  
  d911bc4d
- Created an separate kiss_twiddle_cpx type to make it possible to use · 9ced5d04
  Jean-Marc Valin authored 17 years ago
  
  different precision for twiddles and data.
  9ced5d04
Feb 22, 2008
- MDCT now scales down by N/2 instead of N/4. The factor two is moved to the · 2c1d2f5b
  Jean-Marc Valin authored 17 years ago
  
  overlap-add during synthesis.
  2c1d2f5b
- testcase for the MDCT and IMDCT · b8e5c4c6
  Jean-Marc Valin authored 17 years ago
  
  b8e5c4c6
- Increased precision for real FFT · 5b9c3b35
  Jean-Marc Valin authored 17 years ago
  
  5b9c3b35
- Oops, forgot a file · b8ee0cec
  Jean-Marc Valin authored 17 years ago
  
  b8ee0cec
- Fixed the FFT for higher precision · 25649c15
  Jean-Marc Valin authored 17 years ago
  
  25649c15
- Adding celt_sig_t where needed · 37d13cee
  Jean-Marc Valin authored 17 years ago
  
  37d13cee
- Separating types for normalised vs. denormalised data paths · 18ddc02a
  Jean-Marc Valin authored 17 years ago
  
  18ddc02a
- Fixed the real fft testsuite as well · 14806ab9
  Jean-Marc Valin authored 17 years ago
  
  14806ab9
- Fixed stuff that got broken during the forward-backward split of the FFT · af8402e0
  Jean-Marc Valin authored 17 years ago
  
  af8402e0
- Added a TODO list · df3cb9be
  Jean-Marc Valin authored 17 years ago
  
  df3cb9be
Feb 21, 2008
- Making sure freed or corrupted modes can't be used (produce a run-time warning). · 44ffd5a8
  Jean-Marc Valin authored 17 years ago
  
  44ffd5a8
- Updated the header and celtenc/celtdec tools to the new way modes are handled. · 14f5e7cd
  Jean-Marc Valin authored 17 years ago
  
  14f5e7cd