Commits · 30e1b4d36c432b5211e7be7d0066e91a846ce04d · Alexander Traud / Opus · GitLab

Snippets Groups Projects

Apr 21, 2008
- Moved the windowing operation from compute_mdcts() to mdct_forward() in an · 30e1b4d3
  Jean-Marc Valin authored 16 years ago
  
  attempt to reduce copying.
  30e1b4d3
- Passing the mode to the intra prediction functions so the single-channel case · ba11d783
  Jean-Marc Valin authored 16 years ago
  
  can be optimised
  ba11d783
- Disabling pitch when gain is very small. · 2146e649
  Jean-Marc Valin authored 16 years ago
  
  2146e649
- Making stack push a tiny bit clearer to the compiler -- don't think it really · 431e4646
  Jean-Marc Valin authored 16 years ago
  
  has an effect though.
  431e4646
- s/B/C/ in the intra stuff to have the same notation as everywhere else · 945d0df7
  Jean-Marc Valin authored 16 years ago
  
  945d0df7
- Bit better indexing in intra_prediction() · 8c4877bd
  Jean-Marc Valin authored 16 years ago
  
  8c4877bd
- Removing some unnecessary initialisations to zero. · 50bf8b01
  Jean-Marc Valin authored 16 years ago
  
  50bf8b01
- Another bunch of do-while() loops · dff9b7eb
  Jean-Marc Valin authored 16 years ago
  
  dff9b7eb
Apr 20, 2008
- Further simplifications to comb2pulses() to remove all conditional branches. · d7291d24
  Jean-Marc Valin authored 16 years ago
  
  d7291d24
- Saved some memory by reducing the size of some arrays to only what's needed. · 7aa9d8c7
  Jean-Marc Valin authored 16 years ago
  
  Typically saves 768 bytes.
  7aa9d8c7
- Better indexing of pre/de-emphasis in celt_{en|de}code(). · 62290065
  Jean-Marc Valin authored 16 years ago
  
  62290065
- Saved two copies when computing MDCTs of mono signals · 1a3ed79c
  Jean-Marc Valin authored 16 years ago
  
  1a3ed79c
- more use of CELT_COPY in celt_encode(). Also fixed a stack size bug for stereo · bdb58836
  Jean-Marc Valin authored 16 years ago
  
  bdb58836
- This both reverts 3fe0db17 that caused · 0bc5f7fc
  Jean-Marc Valin authored 16 years ago
  
  accuracy-related issues at high bit-rate, and fixed an overflow caused by large Rxy values.
  0bc5f7fc
Apr 19, 2008
- more restricted pointers in celt.c · 65fb92e8
  Jean-Marc Valin authored 16 years ago
  
  65fb92e8
- added a shortcut (SHORTCUT) to skip the spreading function in the pitch search · 70720a37
  Jean-Marc Valin authored 16 years ago
  
  70720a37
- encoder pre-emphasis now in 16-bits · 91ecf24e
  Jean-Marc Valin authored 16 years ago
  
  91ecf24e
- better indexing in exp_rotation() · 53695789
  Jean-Marc Valin authored 16 years ago
  
  53695789
- better mdct_backward() indexing · 884916fb
  Jean-Marc Valin authored 16 years ago
  
  884916fb
Apr 18, 2008
- optimisation: giving more hints to the compiler about the sizes in · f11d6f49
  Jean-Marc Valin authored 16 years ago
  
  find_spectral_pitch()
  f11d6f49
- a few minor optimisations · d583b246
  Jean-Marc Valin authored 16 years ago
  
  d583b246
- removed unneeded variable shifts from alg_quant() · 3fe0db17
  Jean-Marc Valin authored 16 years ago
  
  3fe0db17
- separated the two passes from interp_bits2pulses() · 784959e8
  Jean-Marc Valin authored 16 years ago
  
  784959e8
- optimisation: removed the shifts from the vq_index() inner loop · 05778451
  Jean-Marc Valin authored 16 years ago
  
  05778451
- optimisation: better indexing/looping in vq_index() · 7adca0e6
  Jean-Marc Valin authored 16 years ago
  
  7adca0e6
- pseudo-stack no longer checks on every function entry whether it has been · f7cec83c
  Jean-Marc Valin authored 16 years ago
  
  allocated
  f7cec83c
- optimisation: one less conditional branch in pulse2comb() · ba829524
  Jean-Marc Valin authored 16 years ago
  
  ba829524
- optimisation: Removed a bunch of conditional branches from comb2pulse() · e6ce0c6a
  Jean-Marc Valin authored 16 years ago
  
  e6ce0c6a
- Saturation in SIG2INT16 using MIN/MAX · abdfc388
  Jean-Marc Valin authored 16 years ago
  
  abdfc388
- Removed implicit 32=>16 conversion (changed to EXTRACT16) · fa147077
  Jean-Marc Valin authored 16 years ago
  
  fa147077
- optimisation: Got rid of the 32-bit mul in find_spectral_pitch() · ed90daef
  Jean-Marc Valin authored 16 years ago
  
  ed90daef
Apr 16, 2008
- minor simplification in alg_quant() · eb1f75d1
  Jean-Marc Valin authored 16 years ago
  
  eb1f75d1
- fixed an issue (lacking parentheses) in the no-op version of BITREV · 309d6420
  Jean-Marc Valin authored 16 years ago
  
  309d6420
- Made twiddle pointer in mdct more explicit · 39aad911
  Jean-Marc Valin authored 16 years ago
  
  39aad911
- optimisation: Making it clear to the compiler that many of the loops in cwrs · 558c50eb
  Jean-Marc Valin authored 16 years ago
  
  need to iterate at least once.
  558c50eb
- More use of CELT_MEMSET · 57e004bf
  Jean-Marc Valin authored 16 years ago
  
  57e004bf
- Make use of CELT_MEMSET() in find_spectral_pitch() · 7f28f768
  Jean-Marc Valin authored 16 years ago
  
  7f28f768
- Fixed incorrect energy calculation in stereo intra prediction · b2b71a58
  Jean-Marc Valin authored 16 years ago
  
  b2b71a58
- optimisation: reworked intra_prediction() so that yy is computed fully only · a66c3208
  Jean-Marc Valin authored 16 years ago
  
  once and then updated for each iteration
  a66c3208
- optimisation: intra_prediction() uses a 16-bit numerator for the search · 0a864645
  Jean-Marc Valin authored 16 years ago
  
  0a864645