shithub: opus

RSS feed

4b000c37 – Jean-Marc Valin <[email protected]> authored on 2011/01/26 15:30
Setting bandE[] to zero after log2Amp when silence=1

8b631f2c – Gregory Maxwell <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/26 15:19
Fixes for silence handling in VBR mode, plus an encoder/decoder desync triggered by silent frames.

e3e2c26d – Jean-Marc Valin <[email protected]> authored on 2011/01/26 08:09
Removing more unused function params

13a7c266 – Jean-Marc Valin <[email protected]> authored on 2011/01/26 05:58
Removes explicit filling of remaining bits with zeros

c39bb8ab – Jean-Marc Valin <[email protected]> authored on 2011/01/26 05:50
Removes unused function parameters

4610d183 – Jean-Marc Valin <[email protected]> authored on 2011/01/26 05:00
Should avoid running log2Amp on uninitialized data

de79c378 – Jean-Marc Valin <[email protected]> authored on 2011/01/26 04:24
Adding a special way to code digital silence in two or more bytes

72e86fa2 – Jean-Marc Valin <[email protected]> authored on 2011/01/25 16:43
Remove // comment

9ce95e0b – Jean-Marc Valin <[email protected]> authored on 2011/01/25 14:12
anti-collapse tuning

72a55477 – Jean-Marc Valin <[email protected]> authored on 2011/01/25 10:58
Oops, fixed the sqrt(2) constant for anti_collapse()

d121260f – Jean-Marc Valin <[email protected]> authored on 2011/01/25 08:11
Minimum period is now 15

01fa3389 – Jean-Marc Valin <[email protected]> authored on 2011/01/24 12:21
Taking LM into account for anti-collapse

68b8d72e – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/24 07:50
Fix off-by-one error in ec_laplace_encode.

495114b7 – Jean-Marc Valin <[email protected]> authored on 2011/01/24 10:53
Moving energy floor to coarse quantization

6ec2ca56 – Jean-Marc Valin <[email protected]> authored on 2011/01/23 13:22
Clipping the encoder input to +/-2 to prevent scaling errors

3a56c9e1 – Jean-Marc Valin <[email protected]> authored on 2011/01/23 06:34
prefilter/postfilter now forced off in Opus hybrid mode

eafd8a7f – Jean-Marc Valin <[email protected]> authored on 2011/01/22 19:24
Simple DTX/CNG implementation

2293e461 – Jean-Marc Valin <[email protected]> authored on 2011/01/22 17:29
Considering "badness" in two-pass quant_coarse_energy().

ad3d813f – Jean-Marc Valin <[email protected]> authored on 2011/01/22 16:22
Stop uselessly calling haar1() when resynth=0

568326bf – Gregory Maxwell <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/22 15:16
Prevent out of bounds access to collapse_masks.

8f02c482 – Gregory Maxwell <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/22 09:57
Correct an encoder/decoder mismatch at low volume levels. Relax some low level clamps so that the dynamic range can extend further below the 16bit floor.

5c2ac2b7 – Jean-Marc Valin <[email protected]> authored on 2011/01/22 09:48
Tracking the background noise level

620e716b – Jean-Marc Valin <[email protected]> authored on 2011/01/21 13:41
Defining merge_stereo() when S==-M or S==M

6bdc53b8 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/20 18:08
Fix collapse_masks overflow for mono.

5da938b2 – Jean-Marc Valin <[email protected]> authored on 2011/01/20 19:07
Reorders some operations in anti-collapse to reuse values

986e2695 – Jean-Marc Valin <[email protected]> authored on 2011/01/20 19:02
Prevents taking the log of zero in fixed-point

63fb61f1 – Jean-Marc Valin <[email protected]> authored on 2011/01/20 17:52
Using previous range coder state for PRNG

e8a373fd – Jean-Marc Valin <[email protected]> authored on 2011/01/19 15:47
Fixes an irrelevant uninitialized bug

a363e395 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/18 05:44
Remove useless ec_dec_tell() call.

21af73eb – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/19 11:30
Make collapse-detection bitexact.

87efe1df – Jean-Marc Valin <[email protected]> authored on 2011/01/18 09:44
Adds an anti-collapse mechanism for transients

2ce5c63d – Jean-Marc Valin <[email protected]> authored on 2011/01/17 15:50
Moving the tapset signalling to the beginning of the stream

8d367029 – Jean-Marc Valin <[email protected]> authored on 2011/01/17 11:37
Adding tapset decision logic

dfa847a2 – Jean-Marc Valin <[email protected]> authored on 2011/01/17 06:37
Support for multiple postfilter tapsets

d85018cb – Gregory Maxwell <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/13 09:38
In CVBR mode the rate selection was failing to add bytes which were about to fall off the end of the bitres and never be reusable, causing undershoot.

5677e34f – Jean-Marc Valin <[email protected]> authored on 2011/01/13 11:15
Setting oldBandE to zero outside of [start,end[

f0d828fc – Jean-Marc Valin <[email protected]> authored on 2011/01/13 10:01
Minor code cleanup, nothing to see here

a387ebfc – Jean-Marc Valin <[email protected]> authored on 2011/01/13 08:22
Proper scaling for the mid folding

173774bb – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/12 11:34
Replace log2_frac in the delta offset calculation.

2b13401f – Jean-Marc Valin <[email protected]> authored on 2011/01/12 11:13
Allowing the tf recombining to go all the way to LM=3

6b565268 – Jean-Marc Valin <[email protected]> authored on 2011/01/12 06:27
Fixes constrained VBR

44203907 – Jean-Marc Valin <[email protected]> authored on 2011/01/12 04:22
Enforces bands of even size even for custom modes

235c64b9 – Jean-Marc Valin <[email protected]> authored on 2011/01/11 12:05
Using a step pdf for the stereo itheta encoding

c52d7689 – Jean-Marc Valin <[email protected]> authored on 2011/01/11 04:42
Minor fixes to testcases

da290c88 – David Schleef <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/01 15:03
Add --export-symbols-regex

b045a26a – David Schleef <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/01 14:52
MSVC build fixes

2d333b4d – David Schleef <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/01 14:47
Use more standard test for lrintf/lrint

ecefde3d – Jean-Marc Valin <[email protected]> authored on 2011/01/11 04:20
Fixes the recombining stride and the deinterleaving stride

8cfda4a3 – Jean-Marc Valin <[email protected]> authored on 2011/01/11 01:39
Using intensity_stereo() when itheta==0

a66b7574 – Jean-Marc Valin <[email protected]> authored on 2011/01/10 08:21
Defines MAX_FINE_BITS to ensure that we're using the same value everywhere

9d2d0510 – Jean-Marc Valin <[email protected]> authored on 2011/01/10 07:45
Using tell() rather than log2_frac() to compute qalloc

d9f6ec3f – Jean-Marc Valin <[email protected]> authored on 2011/01/10 05:52
Changes the N=2 stereo case to use the same sign convention as N=1

c2095a29 – Jean-Marc Valin <[email protected]> authored on 2011/01/10 05:35
Max delta: +/- 16384

d9127edb – Jean-Marc Valin <[email protected]> authored on 2011/01/10 05:34
32-bit fixes

75f99bc5 – Jean-Marc Valin <[email protected]> authored on 2011/01/10 05:02
Make LCG 16-bit clean

08ef1f4c – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/09 15:04
Properly skip padding in testcelt for stereo.

76469c64 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/07 04:18
Prevent busts at low bitrates.

051e044d – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/08 17:11
Fix Jean-Marc's sqrt(0.5) constants.

d0aa9f86 – Jean-Marc Valin <[email protected]> authored on 2011/01/08 20:27
Tuning the split allocation for temporal masking

a714994b – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/04 05:26
Use B0 instead of B for decisions in quant_band().

1cb32aa0 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/04 04:07
Fix rounding in bits2pulses search.

845dfa19 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2011/01/02 11:53
Changes to ec_dec_cdf() to support 8-bit tables.

79d76a2e – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/30 04:06
Code intensity offset relative to start.

d6f61571 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/30 04:04
Make the dynalloc boost run over [start,end).

8893e530 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/30 03:56
Fix the limits for CELT_SET_END_BAND_REQUEST.

fe40ddd9 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/30 03:43
Update the maximum fine bits everywhere.

949f1801 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/30 03:38
Move fine_priority calculation after the bust cap.

fa7215fd – Jean-Marc Valin <[email protected]> authored on 2010/12/29 10:19
Fixing stereo renormalisation

509ad208 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/27 13:20
Move tf_select before the tf_res bits.

41a15e6c – Jean-Marc Valin <[email protected]> authored on 2010/12/27 11:33
Fixes tf_select_table[][] (missing commas)

cf874373 – Jean-Marc Valin <[email protected]> authored on 2010/12/21 20:03
Minor tweaks to the max allocation

20e4c6a6 – Jean-Marc Valin <[email protected]> authored on 2010/12/21 17:27
Removing FIXMEs that were fixed

8fcde379 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/23 06:01
Clean up some type usage.

df2c71ea – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/23 03:06
Use _BitScanReverse for EC_ILOG with MSVC.

1aaa50d1 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/21 12:45
Update ec_dec_cdf() to use an unsigned cdf[].

de31e7e0 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/21 12:30
Add test coverage for entropy coder compatibility.

4e955ed3 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/21 11:36
Fix a typo in ec_byte_write_done().

7365c7d1 – Jean-Marc Valin <[email protected]> authored on 2010/12/21 11:30
Don't enable the post-filter in Opus hybrid mode

30df6cf3 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/21 03:42
Entropy coder clean-up.

59858633 – Jean-Marc Valin <[email protected]> authored on 2010/12/21 06:28
Added highest allocation line possible

dc6d69e6 – Jean-Marc Valin <[email protected]> authored on 2010/12/20 20:04
Ordering Hadamard frequencies when increasing the time-resolution.

034efa52 – Jean-Marc Valin <[email protected]> authored on 2010/12/20 19:19
Tuning the dynamic allocation probability and increment

88bab52a – Jean-Marc Valin <[email protected]> authored on 2010/12/20 11:58
Bump bit-stream version

e8a8582b – Jean-Marc Valin <[email protected]> authored on 2010/12/20 06:40
Bump to 0.10

6ba0b350 – Jean-Marc Valin <[email protected]> authored on 2010/12/20 06:40
Being a bit more careful about enabling the post-filter

6e4b9dc0 – Jean-Marc Valin <[email protected]> authored on 2010/12/19 16:36
Turning noise filling back on for cases where we have no data to fold with

9b98aaa7 – Gregory Maxwell <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/18 21:50
Eliminate some divisions from rate.c.

7c673cfc – Gregory Maxwell <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/18 21:26
Fix for the allocation going negative. (bits[j] >= 0 assert)

53454f49 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/17 23:51
More simplifications to the decoder renormalization loop.

e86fb268 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/17 09:50
Replace ec_{enc|dec}_bit_prob() with ec_{enc|dec}_bit_logp().

8c23a3a0 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/17 09:32
Subtract one from dif in the range decoder.

a0b664df – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/17 05:49
Add a generic CDF decoding routine.

3fed34ae – Jean-Marc Valin <[email protected]> authored on 2010/12/17 09:17
Giving less bits to single-bin bands.

f33a7fb8 – Jean-Marc Valin <[email protected]> authored on 2010/12/17 08:38
Fixed the spreading probabilities (1-x)

320cf2e2 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/17 00:52
Re-organize spreading/folding constants.

cd84e3d0 – Jean-Marc Valin <[email protected]> authored on 2010/12/16 17:29
Re-enabling post-filter on 2.5 ms frames

428a77d6 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/16 11:50
More cleanups to compute_allocation().

76ea41e1 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/16 09:39
Give the bit we reserved to end skipping back when we don't use it.

8e31ab3d – Jean-Marc Valin <[email protected]> authored on 2010/12/16 11:45
Making lower sampling rates use the same allocation table

8e447678 – Gregory Maxwell <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/12/16 09:22
Terminate the coding of skip bits at the last dynalloc boosted band. Otherwise the bitstream allows non-sensible behavior by the encoder (dynallocing bits into a band and then skipping it). This reduces skip bit overhead by about 2-3% at moderate bitrates with the current encoder.

Next