Sign in
android
/
platform
/
external
/
libopus
/
HEAD
« Previous
2aa41de
Remove REAMDE.version
by Harish Mahendrakar
· 5 months ago
3499d0a
switched to smaller NoLACE model
by Jan Buethe
· 5 months ago
ec04a94
bugfix in SilkFeatureNetPL
by Jan Buethe
· 5 months ago
5f8201c
OSCE_MAX_RNN_UNITS now derived from osce model parameters
by Jan Buethe
· 5 months ago
6a9831a
Remove run-time code for old TF2 models
by Jean-Marc Valin
· 5 months ago
1ddfcfd
Using PyTorch model (same architecture for now)
by Jean-Marc Valin
· 5 months ago
e699263
Improving PLC
by Jean-Marc Valin
· 5 months ago
299e38c
Updated LACE and NoLACE models to version 2
by Jan Buethe
· 6 months ago
4f311a1
PLC export script
by Jean-Marc Valin
· 5 months ago
26ddfd7
PyTorch code for training the PLC model
by Jean-Marc Valin
· 5 months ago
6ad03ae
Prevent overshoots from CELT PLC with prediction
by Jean-Marc Valin
· 6 months ago
bd2e9a3
Add simulated loss to opus_demo
by Jean-Marc Valin
· 6 months ago
caca188
Make loss simulator standalone
by Jean-Marc Valin
· 6 months ago
bd710e9
C code for packet loss simulator
by Jean-Marc Valin
· 6 months ago
b923fd1
lossgen: better training, README.md
by Jean-Marc Valin
· 6 months ago
c40add5
lossgen: can now dump weights
by Jean-Marc Valin
· 6 months ago
627aa7f
Packet loss generation model
by Jean-Marc Valin
· 6 months ago
7d328f5
Merge LACE/NoLACE under OSCE framework
by Jan Buethe
· 7 months ago
591c8ba
Initialize padding pointers to zero
by Jean-Marc Valin
· 6 months ago
12fbd81
use opus_(re)alloc and opus_free for dnn and DRED related functions
by Michael Klingbeil
· 6 months ago
f5a1efd
handle extensions in opus_repacketizer_out_range_impl
by Michael Klingbeil
· 6 months ago
6d7ae21
add extensions of the first frame of a multiframe packet
by Michael Klingbeil
· 6 months ago
f27798d
Fix RESYNTH bit rot
by Jean-Marc Valin
· 6 months ago
c7bfc72
use vec_avx.h for MSVC builds
by Michael Klingbeil
· 6 months ago
8090aac
don't redefine _mm_loadu_si32 on MSVC
by Michael Klingbeil
· 6 months ago
88fc293
Defining __SSEx__ macros when needed for MSVC
by Jean-Marc Valin
· 6 months ago
f126bfc
fix autogen.bat model download
by Michael Klingbeil
· 6 months ago
0d823c1
Add a script to shrink the DNN models
by Jean-Marc Valin
· 6 months ago
443510c
Fix Windows path
by Jean-Marc Valin
· 6 months ago
5578824
Fix model download path for windows
by Jean-Marc Valin
· 6 months ago
ddfa480
Opus github ci files
by Jean-Marc Valin
· 6 months ago
08eefed
Add dotprod support to meson
by Jean-Marc Valin
· 6 months ago
c28b0f1
Trying to fix/update meson build
by Jean-Marc Valin
· 7 months ago
147b722
Oops, fix the fixed-point build
by Jean-Marc Valin
· 7 months ago
db26e38
Trying to use fma instructions when possible
by Jean-Marc Valin
· 7 months ago
72cc88d
FARGAN model update
by Jean-Marc Valin
· 7 months ago
df63771
Fixes for ARMv7/AArch32
by Jean-Marc Valin
· 7 months ago
c143b72
Enabling DNN optimizations for ARMv7
by Jean-Marc Valin
· 7 months ago
ee1bb69
Only force auto-vectorization for GCC >= 5.1
by Jean-Marc Valin
· 7 months ago
7cc30ec
Force vectorization for DNN primitives
by Jean-Marc Valin
· 7 months ago
d4506af
Enable floating-point approximations by default
by Jean-Marc Valin
· 7 months ago
db6dad4
Fix ARMv7 optimizations for DNN code
by Jean-Marc Valin
· 7 months ago
cc11c07
First step towards DNN optimization for ARMv7 Neon
by Jean-Marc Valin
· 7 months ago
c9af8f8
Fix potential read out of bounds in fargan
by Jean-Marc Valin
· 7 months ago
5c3795b
Adding dotprod instruction to ARM rtcd
by Jean-Marc Valin
· 7 months ago
984f35b
Speed up cross-correlation normalization
by Jean-Marc Valin
· 7 months ago
d65b7de
Use arch-specific celt_inner_prod() for features
by Jean-Marc Valin
· 7 months ago
ddbdbec
Optimize biquad() to reduce dependency chains
by Jean-Marc Valin
· 7 months ago
176507e
Remove process_single_frame()
by Jean-Marc Valin
· 7 months ago
9d0425d
Remove feature writing (fwrite()) from libopus
by Jean-Marc Valin
· 7 months ago
f582119
Using the same condition for enabling rtcd
by Jean-Marc Valin
· 7 months ago
3e18d96
Trying to fix CMake build
by Jean-Marc Valin
· 7 months ago
239d223
Add rtcd for silk_inner_product_FLP()
by Jean-Marc Valin
· 7 months ago
b93e4a1
Start enabling AVX2 silk_inner_product_FLP()
by Jean-Marc Valin
· 7 months ago
ed90060
Avoids AVX2 optimizations being disabled
by Jean-Marc Valin
· 7 months ago
c066af1
Use SILK VBR when using CBR with DRED
by Jean-Marc Valin
· 7 months ago
6f99a33
Misc fixes on previous patch
by Jean-Marc Valin
· 7 months ago
735c407
Optimize NSQ_del_dec() for AVX2
by Victor Ding
· 7 months ago
452aa95
AVX2 version of silk_inner_product_FLP()
by Jean-Marc Valin
· 7 months ago
1085126
Remove AVX pitch code for fixed-point
by Jean-Marc Valin
· 7 months ago
161358d
Speeding up transient_analysis()
by Jean-Marc Valin
· 7 months ago
f42940b
Make sure weights files are marked as modified
by Jean-Marc Valin
· 7 months ago
d4b04d3
Speed up silk_warped_autocorrelation_FLP()
by Jean-Marc Valin
· 7 months ago
b2cfd87
Add rtcd support for celt_pitch_xcorr_avx2()
by Jean-Marc Valin
· 7 months ago
0293854
Fix non-RTCD case when SSE is not assumed present
by Jean-Marc Valin
· 7 months ago
7423ce5
Use celt_pitch_xcorr_avx2() when guaranteed
by Jean-Marc Valin
· 7 months ago
a93b09e
Adding RTCD for compute_conv2d()
by Jean-Marc Valin
· 7 months ago
91d1f75
FARGAN model update
by Jean-Marc Valin
· 7 months ago
7f7b2a1
Smaller version of fargan
by Jean-Marc Valin
· 7 months ago
19a5d6e
Remove C99 comment
by Jean-Marc Valin
· 7 months ago
4bfc0f8
Adding RTCD for compute_activation()
by Jean-Marc Valin
· 7 months ago
2e034f6
Adding RTCD for DNN code
by Jean-Marc Valin
· 7 months ago
b0620c0
Using sparse GRUs in DRED decoder
by Jean-Marc Valin
· 7 months ago
58923f6
Fix non-AVX builds
by Jean-Marc Valin
· 7 months ago
77594bf
Dumping RDOVAE stats from XML
by Jean-Marc Valin
· 7 months ago
222662d
DRED: quantize scale and dead zone to 8 bits
by Jean-Marc Valin
· 7 months ago
4e10455
added weight export script for LACE/NoLACE
by Jan Buethe
· 7 months ago
8af5c6b
added transposed 1d convolutions to wexchange
by Jan Buethe
· 7 months ago
b6095cf
DRED code cleanup
by Jean-Marc Valin
· 7 months ago
0ab0640
Split stats in two and remove useless dimensions
by Jean-Marc Valin
· 7 months ago
2386a60
updated moc to match results in ietf118 presentation
by Jan Buethe
· 7 months ago
544b3e5
DRED: quantize r and p0 parameters with 8 bits
by Jean-Marc Valin
· 7 months ago
98b8be0
Vectorize DRED quantization
by Jean-Marc Valin
· 7 months ago
fa5e960
Match silenced overflow checks in the sse4.1 version
by Felicia Lim
· 7 months ago
cfeddc4
Silence some overflow checks
by Felicia Lim
· 8 months ago
1ada7d4
Vectorizing sgemv for multiples of 4 with SSE
by Jean-Marc Valin
· 7 months ago
166a6c8
Fix silly bug in CELT Deep PLC
by Jean-Marc Valin
· 7 months ago
74c67a8
Fix CELT PLC for single packet between losses
by Jean-Marc Valin
· 7 months ago
da60266
updated moc method
by Jan Buethe
· 7 months ago
feb3282
Don't try to use models that aren't loaded
by Jean-Marc Valin
· 7 months ago
62b5464
Speed up general case for float matrix multiply
by Jean-Marc Valin
· 7 months ago
61fb3b1
Don't use reserved identifiers for include guards
by Jean-Marc Valin
· 8 months ago
d53531d
Update blob loading code
by Jean-Marc Valin
· 8 months ago
0b75501
Use log approximation when possible
by Jean-Marc Valin
· 8 months ago
4259d35
Reusing already-optimized celt_fir()
by Jean-Marc Valin
· 8 months ago
b22b11a
Silence some warnings
by Jean-Marc Valin
· 8 months ago
ddd5669
Pitch and fargan model updates
by Jean-Marc Valin
· 8 months ago
c99054d
Fix CELT deep PLC bugs
by Jean-Marc Valin
· 8 months ago
ccb244a
cleanup
by Jean-Marc Valin
· 8 months ago
bc102f5
Slightly more continuous analysis
by Jean-Marc Valin
· 8 months ago
Next »