Brad Midgley
d352bd0438
avoid an (unlikely) overflow
2008-01-27 04:29:08 +00:00
Brad Midgley
c2ce7c2d41
get 32-bit products whenever we're sure the multiplicands are both 16 bits
2008-01-27 03:35:53 +00:00
Brad Midgley
c9b5101059
shorten the encoder tables to 16 bits, take out mula32/mul32 for now for simplicity
2008-01-26 19:45:54 +00:00
Brad Midgley
dae52eacba
pcm input array should be 16 not 32 bits
...
use 32-bit product when multiplying two values limited to 16 bits each
2008-01-26 05:24:50 +00:00
Brad Midgley
66fe637352
update copyrights
2008-01-19 15:56:52 +00:00
Brad Midgley
910d620357
coding style
2008-01-14 22:15:17 +00:00
Brad Midgley
8c72b28a0c
comment typo
2008-01-14 20:03:42 +00:00
Brad Midgley
23f3e84ba4
fix initialization
2008-01-14 15:03:23 +00:00
Brad Midgley
d4d085c9a0
take out memmove in sbc analyze
2008-01-14 14:40:57 +00:00
Brad Midgley
0dda8d7230
tweak to the memmove for 4 subbands
2008-01-11 20:28:18 +00:00
Brad Midgley
40d383bb1e
optimizations: use memmove instead of a loop, unroll short loop
2008-01-08 20:56:17 +00:00
Brad Midgley
1da4920f71
smooth out last shift-in-place wrinkle
2007-12-14 06:54:19 +00:00
Brad Midgley
9a14d423e7
push in-place-shift optimization up into scalefactors section
2007-12-14 06:43:06 +00:00
Brad Midgley
c1ce3b25c4
shift-in-place opt is back in, with a bugfix for the 4-subband case
2007-12-14 06:07:52 +00:00
Brad Midgley
5acc7043cf
coding style on ?:
2007-12-14 00:46:09 +00:00
Brad Midgley
356a8b48cd
be more strict about calculating from joint since client may set it to
...
a funky value other than 0/1
2007-12-14 00:15:33 +00:00
Brad Midgley
c5e43eefcc
roll back the shift-in-place bitpack optimization while we figure out if
...
it tickles a bug or creates a bug for 4 subbands
2007-12-14 00:13:07 +00:00
Brad Midgley
2e6b3caf16
don't need to memset the data header
2007-12-13 17:16:20 +00:00
Brad Midgley
87f83cd605
combine two conditionals in bit packing
2007-11-30 15:53:04 +00:00
Brad Midgley
76e595a98b
change inner loop bit packing to use shift-in-place instead of complex ops
2007-11-30 01:20:01 +00:00
Brad Midgley
4161e8201e
don't use the audio_sample array inside the frame struct
...
it's totally unnecessary to keep all these temporary results around
2007-11-29 21:48:12 +00:00
Luiz Augusto von Dentz
56dc26cadc
Remove unnecessary checks for output buffer length and some optimizations.
2007-11-29 20:23:22 +00:00
Brad Midgley
41c4587774
eliminate another max_bitneed loop, eliminate extraneous loop termination tests
2007-11-27 06:47:38 +00:00
Brad Midgley
4a2b8a2e67
eliminate extra loop for bitneed compute
...
make a couple while loops easier to read as for loops
2007-11-27 06:22:48 +00:00
Brad Midgley
6a2bcc32bf
combine the blk loops inside the joint conditional
2007-11-27 03:15:36 +00:00
Brad Midgley
75d93a3225
rework joint subband test--uses a lot less stack space, simplify multiple
...
loops into one, unroll one tiny loop
2007-11-27 02:42:29 +00:00
Luiz Augusto von Dentz
8db224b3ca
Introduce sbc_get_frame_duration.
2007-11-23 14:24:23 +00:00
Luiz Augusto von Dentz
54e6e9fcea
Change name of sbc_fill_defaults to sbc_set_defaults.
2007-11-13 20:59:17 +00:00
Luiz Augusto von Dentz
6d94a5b2e3
Add sbc_reinit.
2007-11-13 20:04:12 +00:00
Luiz Augusto von Dentz
397d6c2b3b
Make sbc codec to write directly in application buffers and so avoiding memcpys.
2007-11-12 18:15:59 +00:00
Brad Midgley
316166212f
correction in the 4 subband optimization
2007-11-05 15:54:50 +00:00
Johan Hedberg
ec233a1f84
More coding style cleanup
2007-11-02 09:05:57 +00:00
Johan Hedberg
9138f99acb
Coding style cleanup
2007-11-01 15:27:38 +00:00
Brad Midgley
f06a303b02
8-subband encoder had a sign error that made high frequencies noisy
2007-10-17 07:36:25 +00:00
Brad Midgley
8fab528b86
a little more factoring
2007-10-06 00:34:35 +00:00
Brad Midgley
3820937249
clean out dead code
2007-10-06 00:01:51 +00:00
Brad Midgley
99b92496ac
fix MUL* macros with ()
...
factor more multiplies in 4-subband encoder
2007-10-05 05:42:55 +00:00
Brad Midgley
7fbcefdd87
enable 4 subband lower half optimizations
2007-10-04 18:28:17 +00:00
Brad Midgley
4cbaa35cb3
fix & enable 8-subband second half optimization
2007-10-04 16:05:40 +00:00
Brad Midgley
aaad5f279f
off-by-one problem in 8 subband optimization
2007-10-04 08:08:59 +00:00
Brad Midgley
6ee3949547
a (disabled for now) cpu optimization for the bottom half of the 8 subbands encoder. seems to work fine.
2007-10-04 07:42:57 +00:00
Brad Midgley
c434c0b495
optimizations for 2nd stage of 4-subband encoder
...
disabled but in place so indt folks don't duplicate the optimization effort
2007-10-04 05:15:07 +00:00
Marcel Holtmann
d9bccae488
Add simple implementation of sbc_parse()
2007-08-26 13:54:24 +00:00
Brad Midgley
c82412ad9e
indicate 0 bytes consumed in encoder if input is too short.
...
caller is responsible for filling the end of the input if it really wants the frame encoded (ie with silence
at the end)
2007-08-26 01:05:25 +00:00
Marcel Holtmann
5689ce1423
Update copyright information
2007-08-25 13:23:51 +00:00
Luiz Augusto von Dentz
b1618922db
Add swap member to sbc struct and fix pcm plugin to not swap the buffer.
2007-08-21 21:32:09 +00:00
Johan Hedberg
46e860574f
Don't hardcode allocation method to SNR
2007-08-11 10:56:46 +00:00
Brad Midgley
a745805cb3
whitespace fixes
2007-03-24 14:01:31 +00:00
Brad Midgley
ad868bd413
update the sbc encoder from the working fixed-point code in the sbc project
...
I also tagged the sbc project with "copied-to-bluez-utils" at the same time.
We will do sbc work under bluez and sync it with the old sbc project if necessary.
2007-03-17 13:37:31 +00:00
Marcel Holtmann
ec31984c45
Add SBC helper includes
2007-02-21 14:23:25 +00:00
