 Jean-Marc Valin
|
e5ad261695
Prevents ridiculously large gains from causing inf/NaNs in float decoder
|
8 years ago |
|
Jean-Marc Valin
|
18a380a7c2
Make it possible to ignore inverted phase stereo for downmix purposes
|
8 years ago |
|
Jean-Marc Valin
|
6db0908db3
Bitstream change: prevent noise injection in hybrid mode
|
8 years ago |
|
Jean-Marc Valin
|
76674feae2
SSE2 implementation of the PVQ search
|
8 years ago |
|
Jean-Marc Valin
|
416611cfa1
biasing quantization
|
8 years ago |
|
Jean-Marc Valin
|
a01feba028
Fixes "mixed declarations and code"
|
8 years ago |
|
Jean-Marc Valin
|
928dcffd98
Weighting theta rdo based on channel energy
|
8 years ago |
|
Jean-Marc Valin
|
abe8b78513
Don't do theta RDO on intensity-stereo-coded bands
|
8 years ago |
|
Jean-Marc Valin
|
eb61c31c14
Saving the state rather than re-compute the best option
|
8 years ago |
|
Jean-Marc Valin
|
35cf18e9d7
Making stereo theta decision based on minimizing distortion
|
9 years ago |
|
Jean-Marc Valin
|
886dc594f0
controlling rounding
|
9 years ago |
|
Jean-Marc Valin
|
3934fac7e2
Properly allocation scratch space for resynth encoder
|
8 years ago |
|
Jean-Marc Valin
|
9eee106214
cleanup: putting resynth flag in the context
|
9 years ago |
|
Jean-Marc Valin
|
3d1b13f507
cleanup: line wrapping
|
9 years ago |
|
Jean-Marc Valin
|
72d89fd7f9
Fixes cap on gain in denormalise_bands()
|
8 years ago |
|
Jean-Marc Valin
|
c76c4de123
Avoid overflowing on ridiculously large energy
|
8 years ago |
|
Jean-Marc Valin
|
b290bca91a
Fixes shift<-2 case for denormalise_bands()
|
8 years ago |
|
Jean-Marc Valin
|
b66080a879
Fixes minor code quality issues in CELT
|
8 years ago |
|
Jean-Marc Valin
|
be9e747bcc
Fixes overflows on CPUs with 16-bit ints
|
8 years ago |
 Jonathan Lennox
|
43120f0075
Reorganize x86 SSE intrinsics code.
|
9 years ago |
 xiangmingzhu
|
c95c9a048f
Cisco optimization for x86 & fixed point
|
10 years ago |
|
Jean-Marc Valin
|
ad7b28ff19
Minor div optimization -- reducing denominators
|
11 years ago |
|
Jean-Marc Valin
|
d9fb8a6651
Optimizing divisions with a signed numerator
|
11 years ago |
|
Jean-Marc Valin
|
29354ff6e0
Save more integer divisions on ARM when we know the operands are positive
|
11 years ago |
|
Jean-Marc Valin
|
bdc7b93358
Reduces decoder stack usage by only storing one channel of denormalized MDCT
|
11 years ago |
|
Jean-Marc Valin
|
4a6744a446
Some cleaning up of the synthesis code.
|
11 years ago |
|
Jean-Marc Valin
|
306d7f5a30
fixed-point: slight (but free) accuracy improvement in compute_band_energies()
|
11 years ago |
|
Jean-Marc Valin
|
e0f26246b0
fixed-point: adds rounding to some shifts to eliminate bias
|
11 years ago |
|
Jean-Marc Valin
|
5752d659fd
Minor fixed-point accuracy improvements that were completely free
|
11 years ago |
|
Jean-Marc Valin
|
5607d5d1c8
Annotating pointer arguments with OPUS_RESTRICT and const
|
11 years ago |
