Commits · ebf72da7604d24d9a5f852ac8a7c9eccb14e5d48 · Mark Harris / Opus

10 Sep, 2008 1 commit
- At least attempting to use all the bits. · ebf72da7
  Jean-Marc Valin authored Sep 09, 2008
  
  ebf72da7
09 Sep, 2008 1 commit
- Simplified allocation by doing it one band at a time, with memory · 825ead82
  Jean-Marc Valin authored Sep 09, 2008
  
  825ead82
01 Sep, 2008 1 commit
- Added Erik de Castro Lopo's float_cast.h to use lrintf() when available. Also · d9b9565a
  Jean-Marc Valin authored Aug 31, 2008
```
disables the synthesis on the encoder side when pitch is disabled.
```
  d9b9565a
30 Aug, 2008 2 commits
- exp() and log() are faster than pow() and log10() for dB conversion · 58b3727f
  Jean-Marc Valin authored Aug 30, 2008
  
  58b3727f
- Added signalling bits for enabling/disabling pitch, short blocks, and folding. · c994394f
  Jean-Marc Valin authored Aug 30, 2008
```
Also, making it possible to disable pitch for all frames.
```
  c994394f
29 Aug, 2008 1 commit
- Fixed a bunch of fixed-point overflows on insanely hot signals by changing · 5c3bc679
  Jean-Marc Valin authored Aug 28, 2008
```
the time-domain representation from Q14 to Q12 (Q29 to Q27 using the
standard convention).
```
  5c3bc679
14 Aug, 2008 1 commit
- Added hooks for tuning pulses and fine energy quant · 8dff923a
  Jean-Marc Valin authored Aug 13, 2008
  
  8dff923a
12 Aug, 2008 1 commit
- Adding a floating point interface to encode and decode · 5f2a74b3
  Gregory Maxwell authored Aug 11, 2008
  
  5f2a74b3
10 Aug, 2008 1 commit
- Fixed fixed-point compilation error. · 8e30e75f
  Gregory Maxwell authored Aug 10, 2008
  
  8e30e75f
07 Aug, 2008 1 commit
- Solved an encoder/decoder mismatch happening when there's no pulse in the first · 5eef264f
  Jean-Marc Valin authored Aug 06, 2008
```
band
```
  5eef264f
06 Aug, 2008 2 commits
- Preventing encoder/decoder mismatch when coarse energy budget is busted · 4117ff02
  Jean-Marc Valin authored Aug 05, 2008
  
  4117ff02
- some tuning to fine energy allocation and prevent it from going above 7 bits · 504f8c25
  Jean-Marc Valin authored Aug 05, 2008
```
(at which point, it causes problems)
```
  504f8c25
04 Aug, 2008 1 commit
- Changing the allocation algorithm to better take into account the fixed cost · 9c70906f
  Jean-Marc Valin authored Aug 03, 2008
```
per frames for each mode.
```
  9c70906f
03 Aug, 2008 3 commits
- Merged compute_allocation_table() and compute_energy_allocation_table() · 36e6e01a
  Jean-Marc Valin authored Aug 02, 2008
  
  36e6e01a
- Turns out that the worse case of the vector split is the same as that of the · 9eba823c
  Jean-Marc Valin authored Aug 02, 2008
```
pulse split. Also, added an allocation line for even higher bit-rates
```
  9eba823c
- fixed a bunch of bugs in the unified allocation code. · eb7ce031
  Jean-Marc Valin authored Aug 02, 2008
  
  eb7ce031
02 Aug, 2008 3 commits
- Fixed two issues pointed out by Greg · 46f3700e
  Jean-Marc Valin authored Aug 02, 2008
  
  46f3700e
- Unified allocation of fine energy and pulses. · 6775de3e
  Jean-Marc Valin authored Aug 02, 2008
  
  6775de3e
- Completed the separation of coarse and fine energy quantisation · c890b58b
  Jean-Marc Valin authored Aug 01, 2008
  
  c890b58b
01 Aug, 2008 1 commit
- Split coarse and fine energy quantisation · 7364e758
  Jean-Marc Valin authored Aug 01, 2008
  
  7364e758
26 Jul, 2008 1 commit
- bump bit-stream version · 400d76e5
  Jean-Marc Valin authored Jul 26, 2008
  
  400d76e5
24 Jul, 2008 1 commit
- Better bit allocation for all frame sizes that are not equal to 256. · 5c0d486f
  Jean-Marc Valin authored Jul 24, 2008
  
  5c0d486f
21 Jul, 2008 3 commits
- Re-implemented renormalise_bands() based on renormalise_vector(). · 1d27ef8b
  Jean-Marc Valin authored Jul 20, 2008
  
  1d27ef8b
- Simplified the folding normalisation by creating renormalise_vector(). · 6361ad8d
  Jean-Marc Valin authored Jul 20, 2008
  
  6361ad8d
- Re-enabling folding/intra for transients · df38f2bf
  Jean-Marc Valin authored Jul 20, 2008
  
  df38f2bf
17 Jul, 2008 1 commit
- isolated the folding part of the code that's common in vq.c · 2c733067
  Jean-Marc Valin authored Jul 17, 2008
  
  2c733067
11 Jul, 2008 1 commit
- Do not encode the short-window bit when not applicable. · 7028d62f
  Jean-Marc Valin authored Jul 10, 2008
  
  7028d62f
06 Jul, 2008 1 commit
- cleaned up transient_analysis() and replaced the algorithm with a simpler one · bd7de0a7
  Jean-Marc Valin authored Jul 06, 2008
```
based on the max absolute value in a range.
```
  bd7de0a7
04 Jul, 2008 2 commits
- fixed-point: conversion of pre-echo avoidance now complete. · f4578196
  Jean-Marc Valin authored Jul 04, 2008
  
  f4578196
- fixed-point: more work on the time window (almost done) · 0ceccb2d
  Jean-Marc Valin authored Jul 03, 2008
  
  0ceccb2d
29 Jun, 2008 1 commit
- More fixed-point conversion of the time window. · e4aeb472
  Jean-Marc Valin authored Jun 29, 2008
  
  e4aeb472
28 Jun, 2008 1 commit
- More fixed-point work on the time window -- including conversion of the gain · 528f4b86
  Jean-Marc Valin authored Jun 29, 2008
```
to power-of-two.
```
  528f4b86
26 Jun, 2008 1 commit
- Cleaned up the pre-echo avoidance code so it works when compiled as fixed-point · 9375aa4f
  Jean-Marc Valin authored Jun 27, 2008
```
(still requires float ops though). Also, removed the per-band IDCT.
```
  9375aa4f
18 Jun, 2008 1 commit
- Automatically choosing the overlap based on the frame size. · 16ca18b9
  Jean-Marc Valin authored Jun 18, 2008
  
  16ca18b9
17 Jun, 2008 1 commit
- short-block MDCT · 9d1decd1
  Jean-Marc Valin authored Jun 17, 2008
  
  9d1decd1
16 Jun, 2008 1 commit
- Implemented two pre-echo avoidance techniques: time-domain pre-emphasis and · 12b22484
  Jean-Marc Valin authored Jun 16, 2008
```
per-band IDCT.
```
  12b22484
11 Jun, 2008 1 commit
- Doing the cwrs split in dimensions should save a few bits. · b0c153b1
  Jean-Marc Valin authored Jun 11, 2008
  
  b0c153b1
10 Jun, 2008 3 commits
- Turns out maxK and maxM were also inverted in fits_in64() · f892d5f6
  Jean-Marc Valin authored Jun 10, 2008
  
  f892d5f6
- fixed a cwrs bug in fits_in32() · 51891c9d
  Jean-Marc Valin authored Jun 10, 2008
  
  51891c9d
- Implemented split-cwrs for very large codebooks (>64 bits), but still getting · 679083f4
  Jean-Marc Valin authored Jun 10, 2008
```
a few decoding errors.
```
  679083f4