shithub: opus

54fb7e56 – Jean-Marc Valin <[email protected]> authored on 2010/10/15 07:26

Tuning the bias for the L1 tf metric

8852563a – Jean-Marc Valin <[email protected]> authored on 2010/10/15 06:24

Unb0rked the VBR code: several units issues that made VBR completely useless

8200b2d1 – Jean-Marc Valin <[email protected]> authored on 2010/10/14 22:18

Modifying the L1 metric for tf analysis by merging together all frequencies.

4a8c1f11 – Jean-Marc Valin <[email protected]> authored on 2010/10/13 14:03

Making it possible to have two transients in a row for longer frames

2794b630 – Jean-Marc Valin <[email protected]> authored on 2010/10/13 13:32

Adding a high-pass filter in front of the transient analysis code

ccd51fe9 – Jean-Marc Valin <[email protected]> authored on 2010/10/13 07:48

No longer filling the stereo side with noise.

ccd5a61a – Jean-Marc Valin <[email protected]> authored on 2010/10/12 21:11

Taking into account TF decisions for VBR

0f8fc0be – Jean-Marc Valin <[email protected]> authored on 2010/10/12 19:25

New TF decision code based on L1-norm. Needs more work.

844f00de – Jean-Marc Valin <[email protected]> authored on 2010/10/12 18:18

Don't take into account silence for the coarse energy means

a8160dde – Jean-Marc Valin <[email protected]> authored on 2010/10/12 10:55

Remove lower bound on bit-rate for tf signalling

f1fea66b – Jean-Marc Valin <[email protected]> authored on 2010/10/09 18:58

Dynamic allocation before VBR

8226ac08 – Jean-Marc Valin <[email protected]> authored on 2010/10/09 17:07

Retrained coarse energy mean and beta coefficients

8ccda887 – Jean-Marc Valin <[email protected]> authored on 2010/10/04 14:01

Making the allocation slightly less insane at very high rate

b441a157 – Jean-Marc Valin <[email protected]> authored on 2010/10/04 10:45

Using high-band noise less aggressively

6bf04627 – Jean-Marc Valin <[email protected]> authored on 2010/09/30 06:16

Allowing to change the allocation dynamically.

4a11daa2 – Gregory Maxwell <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/09/29 14:56

Updates tandem-test for more robust error reporting and additional frame sizes and sample rates.

83c26fbd – Jean-Marc Valin <[email protected]> authored on 2010/09/29 16:05

Fixes a ibudget bust and a divide-by-zero at very low bit-rate

b1fe4f60 – Jean-Marc Valin <[email protected]> authored on 2010/09/29 14:05

Fixes a fixed-point overflow on 16-bit architectures

d74c851f – Jean-Marc Valin <[email protected]> authored on 2010/09/29 13:39

Adds a simple dynamic allocation scheme

b801da5e – Jean-Marc Valin <[email protected]> authored on 2010/09/28 10:56

Preventing bands from being coded at a rate below (for now) 3/8 bit/sample

78ea9fd0 – Jean-Marc Valin <[email protected]> authored on 2010/09/24 04:27

Include tuning from exp_vorbis_tuning (disabled for now)

8abc6bee – Jean-Marc Valin <[email protected]> authored on 2010/09/15 02:57

Fixes a minor folding bug for N=2 bands in stereo

cb2c77c6 – Jean-Marc Valin <[email protected]> authored on 2010/09/14 12:43

Add an extra safety bit in intra_decision()

ba760bc5 – Jean-Marc Valin <[email protected]> authored on 2010/09/14 12:06

Fixes rare overflow in intra_decision()

3c15659c – Jean-Marc Valin <[email protected]> authored on 2010/09/14 12:02

Fixes fixed-point overflow in stereo_merge()

d60e0db0 – Jean-Marc Valin <[email protected]> authored on 2010/09/14 08:22

Fixes a fixed-point overflow in haar1()

9ce95056 – Jean-Marc Valin <[email protected]> authored on 2010/09/14 07:05

Speed up the bit allocator interpolation

5f3ad5f9 – Jean-Marc Valin <[email protected]> authored on 2010/09/14 06:30

Fix stereo for N=2

79b45338 – Jean-Marc Valin <[email protected]> authored on 2010/09/14 05:48

Copy lowband when needed instead of undoing the transforms

ad7d5e00 – Jean-Marc Valin <[email protected]> authored on 2010/09/13 13:44

Allowing a max of 8 bits to qtheta

a5ce03fe – Jean-Marc Valin <[email protected]> authored on 2010/09/13 11:06

More quant_band() reorganisation.

a18c75e5 – Jean-Marc Valin <[email protected]> authored on 2010/09/13 10:59

Restructuring in quant_band()

293b9f19 – Jean-Marc Valin <[email protected]> authored on 2010/09/13 10:53

Fixing tests on machines with no __builtin_clz()

35780760 – Jean-Marc Valin <[email protected]> authored on 2010/09/13 09:56

Separates stereo_band_mix() into the intensity and MS stereo cases

90e7fdd7 – Jean-Marc Valin <[email protected]> authored on 2010/09/13 09:44

Added stereo_merge(), which does the renormalisation too

80cae178 – John Ridges <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/09/13 07:23

MacOS X typedef fix

009978ee – Jean-Marc Valin <[email protected]> authored on 2010/09/13 07:05

Moves the bit-side gain application to the quantizer

ae319fe6 – Jean-Marc Valin <[email protected]> authored on 2010/09/13 05:46

Using 6 steps in allocation interpolation

eea914cb – Jean-Marc Valin <[email protected]> authored on 2010/09/12 16:11

Simplifies vector renormalisation (and using it less)

725e5f4b – Jean-Marc Valin <[email protected]> authored on 2010/09/11 04:57

Fix a minor, but bitstream-affecting bug

5277d2db – Jean-Marc Valin <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/09/10 13:40

Fixes a divide-by-zero in narrowband mode with small frame size

aaedf171 – Jean-Marc Valin <[email protected]> authored on 2010/09/10 16:20

Simplifies convoluted raw bits calls.

f9fdbffb – Jean-Marc Valin <[email protected]> authored on 2010/09/05 17:02

Don't allow transients for 2.5 ms frames.

9e72ae21 – Jean-Marc Valin <[email protected]> authored on 2010/09/03 11:12

Fixes an index bug in dump_modes

5c559ede – Jean-Marc Valin <[email protected]> authored on 2010/09/03 10:27

More sharing across different static modes

e0c25452 – Jean-Marc Valin <[email protected]> authored on 2010/09/03 07:52

static modes simplifications for std mode

5493ded3 – Jean-Marc Valin <[email protected]> authored on 2010/09/02 06:50

minor tuning in the mid-high rates

cef1d6ac – Jean-Marc Valin <[email protected]> authored on 2010/09/02 06:16

Raw bit init for for architectures with chars larger than 8 bits

7fde0588 – Jean-Marc Valin <[email protected]> authored on 2010/09/01 18:04

Smoothing out some curves

7d1cc9d5 – Jean-Marc Valin <[email protected]> authored on 2010/09/01 12:21

Tuning around 30 kb/s

fbcec9a9 – Jean-Marc Valin <[email protected]> authored on 2010/09/01 12:09

Tuning in the 24 kb/s area

c98ff908 – Jean-Marc Valin <[email protected]> authored on 2010/09/01 11:43

Tuning at very low bit-rate (around 14 kb/s)

8bcd3882 – Jean-Marc Valin <[email protected]> authored on 2010/09/01 07:11

Tuning at very low rate

8f50b0f7 – Jean-Marc Valin <[email protected]> authored on 2010/09/01 07:00

Making the allocation matrix a bit smoother

1887401c – Jean-Marc Valin <[email protected]> authored on 2010/09/01 06:07

Converting allocation table to 1/32 bit/sample resolution.

1b36d6c8 – Jean-Marc Valin <[email protected]> authored on 2010/08/31 13:21

Complexity setting now controls two-pass coarse energy quantization

020ef4f8 – Jean-Marc Valin <[email protected]> authored on 2010/08/31 13:08

Ignoring lower bands in folding_decision()

1213ba5e – Jean-Marc Valin <[email protected]> authored on 2010/08/31 13:03

Complexity setting now does something again

a6a53ab5 – Jean-Marc Valin <[email protected]> authored on 2010/08/31 12:46

Minor simplification to interp_bits2pulses()

1d17b9ae – Jean-Marc Valin <[email protected]> authored on 2010/08/31 10:51

New implementation of folding_decision()

6f6c88be – Jean-Marc Valin <[email protected]> authored on 2010/08/31 07:56

_alloca() for Win32

736efd69 – Jean-Marc Valin <[email protected]> authored on 2010/08/31 07:52

Fixes some MSVC warnings

5c0c936d – Jean-Marc Valin <[email protected]> authored on 2010/08/31 06:11

Fine energy allocation cleanup

f1c8fb10 – Jean-Marc Valin <[email protected]> authored on 2010/08/31 02:45

qoffset tuning

32a1e6b6 – Jean-Marc Valin <[email protected]> authored on 2010/08/30 19:35

More fine energy tuning, compensation for N=2

fffae794 – Jean-Marc Valin <[email protected]> authored on 2010/08/30 17:34

New fine energy allocation tuning.

6d3829f2 – Jean-Marc Valin <[email protected]> authored on 2010/08/27 13:52

CELT no longer prints to stderr (unless assertions are enabled)

c09807d3 – Jean-Marc Valin <[email protected]> authored on 2010/08/27 13:17

Simplifies the implementation of RESET by placing all the data that needs

7cfb7305 – Jean-Marc Valin <[email protected]> authored on 2010/08/27 12:54

Adds _init() functions that do not allocate the states themselves

ca8b9928 – Jean-Marc Valin <[email protected]> authored on 2010/08/27 12:23

Comments on encoder and decoder struct contents

f9950e06 – Jean-Marc Valin <[email protected]> authored on 2010/08/27 12:15

Remove model markers

bce1dd0d – Jean-Marc Valin <[email protected]> authored on 2010/08/27 12:10

Remove check_mode()

d13cd156 – Jean-Marc Valin <[email protected]> authored on 2010/08/27 12:05

Unb0rk fixed-point build

0410e9ca – Jean-Marc Valin <[email protected]> authored on 2010/08/27 12:02

Removing check_decoder()

02a3527a – Jean-Marc Valin <[email protected]> authored on 2010/08/27 12:00

Decoder state stored in a single allocated object.

6d131087 – Jean-Marc Valin <[email protected]> authored on 2010/08/27 11:15

Decoder cleanup

190b782e – Jean-Marc Valin <[email protected]> authored on 2010/08/27 11:03

Removing check_encoder()

0d8eaa33 – Jean-Marc Valin <[email protected]> authored on 2010/08/27 10:57

Encoder state now stored in a single allocated object

0695a5fb – Jean-Marc Valin <[email protected]> authored on 2010/08/27 07:33

Simplifies convoluted/idiotic indexing.

f67b4476 – Jean-Marc Valin <[email protected]> authored on 2010/08/26 21:32

More IMDCT de-uglification

e0ce2cf9 – Jean-Marc Valin <[email protected]> authored on 2010/08/26 19:58

Removing out_mem from the encoder state.

903dbf75 – Jean-Marc Valin <[email protected]> authored on 2010/08/26 16:06

Inverse MDCT no longer keeps channels interleaved.

ab3784e9 – Jean-Marc Valin <[email protected]> authored on 2010/08/26 14:08

Small suckage reduction in compute_inv_mdcts(). More to come.

2cc1fcbd – Jean-Marc Valin <[email protected]> authored on 2010/08/26 04:17

Nothing to see here

5d937c02 – Jean-Marc Valin <[email protected]> authored on 2010/08/25 22:04

Disabling mdct and fft init code with static modes

40603b2f – Jean-Marc Valin <[email protected]> authored on 2010/08/25 19:02

No longer making copies of static modes

ae035a54 – Jean-Marc Valin <[email protected]> authored on 2010/08/25 18:41

Fixes an off-by-one bug in the static mode's MDCT twiddles

0ff9be6a – Jean-Marc Valin <[email protected]> authored on 2010/08/25 18:20

Fixed an old issue where the fixed-point MDCT twiddles were stored as 32-bit values

bb528817 – Jean-Marc Valin <[email protected]> authored on 2010/08/25 18:12

coarse probability model in static modes too

31b189b4 – Jean-Marc Valin <[email protected]> authored on 2010/08/25 17:21

static modes update for MDCT

24eef149 – Jean-Marc Valin <[email protected]> authored on 2010/08/25 14:01

Updating dump_modes to include the MDCT and FFT. More work needed.

3fc0aada – Jean-Marc Valin <[email protected]> authored on 2010/08/25 13:15

FFT cleanup

732ea388 – Jean-Marc Valin <[email protected]> authored on 2010/08/25 09:52

Updated static modes for new pulse cache.

3ad8db49 – Jean-Marc Valin <[email protected]> authored on 2010/08/25 09:11

New pulse cache

ba3a2892 – Jean-Marc Valin <[email protected]> authored on 2010/08/23 07:14

MSE measurement details (no codec changes)

3f9857b9 – Jean-Marc Valin <[email protected]> authored on 2010/08/13 17:20

Making the fine energy allocation code less ugly.

c1c40a76 – Timothy B. Terriberry <[email protected]> authored and Jean-Marc Valin <[email protected]> committed on 2010/08/11 04:35

Rework coarse energy coding.

6ac26ecf – Jean-Marc Valin <[email protected]> authored on 2010/08/11 16:38

Fixes a stereo bug introduced in the previous commit

bb338a93 – Jean-Marc Valin <[email protected]> authored on 2010/08/10 09:20

Choosing intra frame energy when it's cheaper than inter

5e7f02de – Jean-Marc Valin <[email protected]> authored on 2010/08/08 05:48

Moving intra decision to quant_coarse_energy()