Commits · c163fc62075d5455f3d233af8c291bada6a2c2cd · Alexander Traud / Opus · GitLab

Snippets Groups Projects

Mar 21, 2008
- local var name maxval was shadowing the TI function used to compute it · c163fc62
  Jean-Marc Valin authored 17 years ago
  
  c163fc62
- make sure TRIG_UPSCALE is properly defined · 03de605e
  Jean-Marc Valin authored 17 years ago
  
  03de605e
- fix for TI version of celt_maxabs16() · 59f42b5d
  Jean-Marc Valin authored 17 years ago
  
  59f42b5d
- fixed-point: fix for 32-bit TI FFT · 6e09f680
  Jean-Marc Valin authored 17 years ago
  
  6e09f680
- fixed-point: Wrapper for the 32-bit complex FFT used in the MDCT so we can use · 70162a4f
  Jean-Marc Valin authored 17 years ago
  
  the TI dsplib FFT.
  70162a4f
- fixed-point: defined celt_maxabs16() as basic operator · 9901cb9e
  Jean-Marc Valin authored 17 years ago
  
  9901cb9e
Mar 20, 2008
- fixed-point: MULT16_32_Q15 for TI DSP (not entirely happy with it) · 948dabc7
  Jean-Marc Valin authored 17 years ago
  
  948dabc7
- s/ROUND/ROUND16/ · f5b0587b
  Jean-Marc Valin authored 17 years ago
  
  f5b0587b
- fixed-point: using TI intrinsic for celt_ilog2() if available. · 83006eec
  Jean-Marc Valin authored 17 years ago
  
  83006eec
- Wrapper for the TI dsplib FFT · 983f6387
  Jean-Marc Valin authored 17 years ago
  
  983f6387
- Making the real/single FFT easier to replace · fcb841aa
  Jean-Marc Valin authored 17 years ago
  
  fcb841aa
Mar 18, 2008
- Random numbers should work on 16-bit archs. · 104c218c
  Jean-Marc Valin authored 17 years ago
  
  104c218c
Mar 16, 2008
- fixed-point: more TI macros. Comments on the existing ones. · b311554c
  Jean-Marc Valin authored 17 years ago
  
  b311554c
- Removed useless functions that caused warnings in skeleton.c · b674a278
  Jean-Marc Valin authored 17 years ago
  
  b674a278
Mar 15, 2008
- Making sure not to use the C library calls directly · c7e0b76c
  Jean-Marc Valin authored 17 years ago
  
  c7e0b76c
- Some work on assertions. · 4ff068e6
  Jean-Marc Valin authored 17 years ago
  
  4ff068e6
Mar 14, 2008
- New C55 macro · bfcbd184
  Jean-Marc Valin authored 17 years ago
  
  bfcbd184
- fixed-point: Fixed a "shift too large" warning by using a 16-bit mul instead · fb7359a6
  Jean-Marc Valin authored 17 years ago
  
  of 32-bit in find_spectral_pitch()
  fb7359a6
- fixed-point: playing it safe. SHL32() now automatically casts input to 32-bit · 15b9be3a
  Jean-Marc Valin authored 17 years ago
  
  to prevent surprises. Also, a few comments.
  15b9be3a
- fixed-point: fixed a 16-bit issue in normalise_bands() where Q15ONE was directly · 0773fec4
  Jean-Marc Valin authored 17 years ago
  
  shifted left
  0773fec4
- Removed unnecessary header inclusions · 71a84590
  Jean-Marc Valin authored 17 years ago
  
  71a84590
Mar 13, 2008
- Adjusting/fixing warnings · a82dfdd2
  Jean-Marc Valin authored 17 years ago
  
  a82dfdd2
- fixed-point: Using reciproqual instead of DIV32_16 for find_spectral_pitch() · 2b35de91
  Jean-Marc Valin authored 17 years ago
  
  2b35de91
- Enabling warnings and restricting symbol visibility · 679f5cc6
  Jean-Marc Valin authored 17 years ago
  
  679f5cc6
- Added mathops-test · 92518982
  Jean-Marc Valin authored 17 years ago
  
  92518982
- fixed-point: simplification of the gain in mix_pitch_and_residual() · 9d5b4a6f
  Jean-Marc Valin authored 17 years ago
  
  9d5b4a6f
- fixed-point: simplifying the arithmetic in alg_quant() · 19ae9fc9
  Jean-Marc Valin authored 17 years ago
  
  19ae9fc9
Mar 12, 2008
- More accurate sqrt approximation using MULT16_16_Q15() instead of Q14. · 3ec78b17
  Jean-Marc Valin authored 17 years ago
  
  3ec78b17
- TODO update · f823f409
  Jean-Marc Valin authored 17 years ago
  
  f823f409
- Increased accuracy of the reciprocal approximation and made it use · 887ad9ff
  Jean-Marc Valin authored 17 years ago
  
  MULT16_16_Q15() instead of Q14 (should be faster).
  887ad9ff
- Testcases should now work even when symbols aren't visible in the dso · 0ac437b8
  Jean-Marc Valin authored 17 years ago
  
  0ac437b8
- Added info for properly exporting symbols · 7f38906f
  Jean-Marc Valin authored 17 years ago
  
  7f38906f
- Changed definition of VARDECL to make it more flexible. No actual code change. · 31b79d19
  Jean-Marc Valin authored 17 years ago
  
  31b79d19
- Using reciprocal approximation instead of full 32-bit division in alg_quant() · d857ac48
  Jean-Marc Valin authored 17 years ago
  
  d857ac48
- configure update (MIXED_PRECISION and STATIC_MODES can now be used) · 980ad384
  Jean-Marc Valin authored 17 years ago
  
  980ad384
- Defined the celt_rcp() reciprocal approximation · 931db72e
  Jean-Marc Valin authored 17 years ago
  
  931db72e
Mar 11, 2008
- __GNUC_PREREQ should be causing problems anymore. · 3c7ae6c3
  Jean-Marc Valin authored 17 years ago
  
  3c7ae6c3
- Added macro definitions for the TI C5x family (untested) · 72e8003f
  Jean-Marc Valin authored 17 years ago
  
  72e8003f
- Better heuristic for deciding when to use the 32-bit version of the cwrs code. · 493b38d4
  Jean-Marc Valin authored 17 years ago
  
  493b38d4
Mar 10, 2008
- const correctness. celt_mode_create() now makes a copy of the static mode · 949902fd
  Jean-Marc Valin authored 17 years ago
  
  (if STATIC_MODES is defined) instead of attempting to modify read-only storage.
  949902fd