RSS feed
f81a60ca
– Jean-Marc Valin <
[email protected]>
authored
on 2010/08/03 11:47
DOUBLE_PRECISION and MIXED_PRECISION no longer need to be defined
bc272de7
– Jean-Marc Valin <
[email protected]>
authored
on 2010/08/02 05:41
Code simplifications for log->amplitude conversion
531f2ae7
– Jean-Marc Valin <
[email protected]>
authored
on 2010/08/02 05:01
Remove useless use of "long", remove useless prototypes
7b5a086b
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/29 11:01
Measuring the normalized error directly within the encoder
8d186014
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/28 18:12
Fixing the qtheta dependency for the delta allocation
ea17d423
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/28 06:39
Removing uselessly low bitrate allocation entries
663c0492
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/24 17:36
Spreading code cleanup -- now allowing tf change when spreading is off
f276b234
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/24 13:13
Using random noise in upper bands when signal is "normal"
1b0bc2d2
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/24 03:59
Updated static mode generation for latest changes
bdfb73bc
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/23 13:46
Just removing old code that was commented out anyway
2ed05631
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/20 13:30
Fixes a bug that could turn off folding at low rate when specifying
6bf3b0a7
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/19 10:32
The coarse energy budget is no longer part of the bit-stream.
9d785afb
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/18 05:42
Encoder now has a way to check whether an error has occurred
8952c45e
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/16 17:48
The encoder and decoder can now process audio encoded at a higher rate
85f41b20
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/16 14:12
Some work towards being able to encode a 48 kHz stream from 32 kHz audio (incomplete)
3b0df0dc
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/16 11:55
All modes based on 2.5 ms short blocks now use the same definition
69921c4c
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/16 11:25
Considering the end band in more places in the code
0be05b03
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/16 10:23
Encoding transient_time independently of the sampling rate
a2546bbf
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/16 08:00
Number of short blocks now based on the frame duration instead of the
af1fce93
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/16 07:05
Pre-emphasis coefficients now part of the mode and has a numerator as well
e34c85a0
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/15 13:57
Using short blocks a bit more often now that we have variable tf resolution
2ed5e67a
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/13 12:50
Handling the start band and end band in tf encoding
2b5a2e7b
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/08 10:35
Fix accidental change of celt_alloc() to non-inline
9310096e
– Jean-Marc Valin <
[email protected]>
authored
on 2010/07/04 20:33
Fix misleading comments about ec_{enc|dec}_tell()
a0110658
– Jean-Marc Valin <
[email protected]>
authored
on 2010/06/23 11:22
Improved quality of small frame sizes at low bitrate.
c51e98ba
– Jean-Marc Valin <
[email protected]>
authored
on 2010/06/03 18:11
Base everything on the "standard allocation table"
12e851da
– Jean-Marc Valin <
[email protected]>
authored
on 2010/06/03 04:12
Using allocation table for "standard" frame sizes
43e94062
– Timothy B. Terriberry <
[email protected]>
authored
and Jean-Marc Valin <
[email protected]>
committed
on 2010/05/29 19:02
Change ec_{enc|dec}_bit_prob to take probabilities in units of 1/65536 instead of 1/256. This allows them to use a single 16x16->32 multiply instead of a 24x8->32 multiply. Also change the time-frequency resolution flag coding to ensure that "0" is always the most-probable symbol (i.e., that prob("1")<50%), as that's where all the rounding error accumulates.
Next