Gitiles
Code Review
Sign In
gerrit-public.fairphone.software
/
platform
/
external
/
libopus
/
d70b8ed69a7761fb2feb5df84e708e638f4e066c
/
celt
c232f65
Adds "arch" argument to MIPS MDCT
by Rhishikesh Agashe
· 9 years ago
db7a0d6
Fix dump_modes output for fixed-point NE10.
by Timothy B. Terriberry
· 9 years ago
bae8545
armv7: Enable NE10's fixed-point FFT for the MDCT
by Viswanath Puttagunta
· 9 years ago
ba8713c
armv7: Optimize fixed point FFT using NE10 library
by Viswanath Puttagunta
· 9 years ago
3807af3
test_unit_dft: Add nfft = 60, 240, 480 tests
by Viswanath Puttagunta
· 9 years ago
19c5406
armv7(float): Optimize decode usecase using NE10 library
by Viswanath Puttagunta
· 9 years ago
f48abe8
armv7(float): Optimize encode usecase using NE10 library
by Viswanath Puttagunta
· 9 years ago
d71c297
Add entcode.c to the dump_modes binary.
by Timothy B. Terriberry
· 9 years ago
c9c2035
Fixes compile problems for MIPS
by Rhishikesh Agashe
· 9 years ago
fee0ed0
Add intrinsics support to Visual Studio build.
by Jonathan Lennox
· 9 years ago
43120f0
Reorganize x86 SSE intrinsics code.
by Jonathan Lennox
· 9 years ago
cb0875e
Move SSE2 and SSE4.1 intrinsics functions to separate files, to be compiled with appropriate compiler flags. Otherwise, compilers are allowed to take advantage of (e.g.) -msse4.1 to generate code that uses SSE4.1 instructions, even when no SSE4.1 intrinsics are explicitly used in the source.
by Jonathan Lennox
· 9 years ago
26fc5c3
Remove some unnecessary #includes from x86cpu.c.
by Jonathan Lennox
· 9 years ago
577bfb1
Fix struct initialization of CPU_Feature structure.
by Jonathan Lennox
· 9 years ago
33cf9e2
Fix cpuid asm on 32-bit PIC.
by Jonathan Lennox
· 9 years ago
1d60b49
In optimized mode, don't force Clang to use explicit load/store for _mm_cvtepi16_epi32, only for _mm_cvtepi8_epi32. Adjust comment accordingly.
by Jonathan Lennox
· 9 years ago
b4aa5dc
Reorganize configure's detection of intrinsics functions:
by Jonathan Lennox
· 9 years ago
2d83e7e
Wrap _mm_cvtepi...() intrinsics in macros on clang.
by Timothy B. Terriberry
· 9 years ago
75d81f5
Document how to tell if your build is fixed-point.
by Timothy B. Terriberry
· 9 years ago
7422189
Fix silk_VQ_WMat_EC_sse4_1().
by Timothy B. Terriberry
· 10 years ago
23f503a
Fix warnings in pedantic build.
by Timothy B. Terriberry
· 10 years ago
0b0a2b4
armv7: celt_pitch_xcorr: Introduce ARM NEON intrinsics
by Viswanath Puttagunta
· 10 years ago
aad2818
Fix celt_pitch_xcorr_c signature.
by Timothy B. Terriberry
· 10 years ago
8921f82
float_cast: Fix MSVC ARM build
by Cameron Gutman
· 10 years ago
fd562ee
tests: fix linking for --enable-fixed-point on ARM
by Tristan Matthews
· 10 years ago
554b349
There are no tabs in source code.
by Timothy B. Terriberry
· 10 years ago
020a0a4
Fix build bustage from c95c9a048.
by Timothy B. Terriberry
· 10 years ago
c95c9a0
Cisco optimization for x86 & fixed point
by xiangmingzhu
· 10 years ago
dacfd28
os_support: fix misleading comments
by Tristan Matthews
· 10 years ago
62351fa
Whitespace fixes
by Jean-Marc Valin
· 10 years ago
f133bac
MIPS optimizations
by Rhishikesh Agashe
· 10 years ago
f396c08
Adds SIG2WORD16() to fixed_debug.h
by Jean-Marc Valin
· 10 years ago
6d2d5c4
Suppress some coverity false positives.
by Gregory Maxwell
· 10 years ago
4ae4bc6
Fix iOS builds with assembly.
by Timothy B. Terriberry
· 10 years ago
76e831d
Make the arm2gnu.pl converter handle apple specific details
by Martin Storsjo
· 10 years ago
ac0e294
Add separate labels for the start of public functions
by Martin Storsjo
· 10 years ago
e70faf9
arm: Use the UAL syntax for instructions
by Martin Storsjo
· 10 years ago
f17623f
Fixes an aliasing bug in the MDCT when the frame size isn't a multiple of 4.
by Marcello Caramma (mcaramma)
· 10 years ago
718b10c
Adds missing license for arm2gnu.pl
by Jean-Marc Valin
· 10 years ago
e318744
Skips comb filter overlap when the parameters didn't change.
by Jean-Marc Valin
· 10 years ago
ad7b28f
Minor div optimization -- reducing denominators
by Jean-Marc Valin
· 10 years ago
d9fb8a6
Optimizing divisions with a signed numerator
by Jean-Marc Valin
· 10 years ago
5123175
Speeding up ec_tell_frac()
by Jean-Marc Valin
· 10 years ago
c543ec1
Speeding up extract_collapse_mask() slightly
by Jean-Marc Valin
· 11 years ago
2e6876c
Using SSAT in SIG2WORD16() on ARMv6
by Jean-Marc Valin
· 11 years ago
9b3a3ad
Unrolled version of the comb filter for ARM (eliminates MOVs)
by Jean-Marc Valin
· 11 years ago
6218cba
Making decode_pulses() compute the L2-norm on the fly
by Jean-Marc Valin
· 11 years ago
17b1978
Speed up the comb filter on ARM by using MAC16_32_Q16()
by Jean-Marc Valin
· 11 years ago
29354ff
Save more integer divisions on ARM when we know the operands are positive
by Jean-Marc Valin
· 11 years ago
ce1173c
Fixes use of uninitialized values in dynalloc_analysis()
by Jean-Marc Valin
· 11 years ago
ec5d01c
Using a table on ARM for unsigned division by small (<=256) integers.
by Jean-Marc Valin
· 11 years ago
379af35
Minor cleanup in dynalloc_analysis()
by Jean-Marc Valin
· 11 years ago
05548fa
Adds a median filter to make dynalloc_analysis() more conservative
by Jean-Marc Valin
· 11 years ago
a65db56
Fix declaration after statement in fixed point.
by Gregory Maxwell
· 11 years ago
e775090
pseudostack instrumentation (off by default)
by Jean-Marc Valin
· 11 years ago
9134e96
Fixes SMALL_FOOTPRINT for float
by Jean-Marc Valin
· 11 years ago
e17ca25
Don't allocate pulses on the stack when calling the SILK PLC.
by Jean-Marc Valin
· 11 years ago
b63e711
Moves CELT PLC pitch search to a separate function to reduce peak stack
by Jean-Marc Valin
· 11 years ago
5f807c1
Adds SMALL_FOOTPRINT hack to the CELT PLC too
by Jean-Marc Valin
· 11 years ago
9d1b6fe
Moves deemphasis() call out of celt_decode_lost() to reduce peak stack
by Jean-Marc Valin
· 11 years ago
ad8371d
Cleaning up leftovers of "freq" in celt_decode_with_ec()
by Jean-Marc Valin
· 11 years ago
4d07b13
Reduces the decoder stack use by removing the pcm_silk buffer in fixed-point
by Jean-Marc Valin
· 11 years ago
14ca4ed
Moves the remains of compute_inv_mdcts() to celt_synthesis()
by Jean-Marc Valin
· 11 years ago
32454dc
Hack that makes the SMALL_FOOTPRINT CELT decoder use only 4.25 kB of stack.
by Jean-Marc Valin
· 11 years ago
bdc7b93
Reduces decoder stack usage by only storing one channel of denormalized MDCT
by Jean-Marc Valin
· 11 years ago
4a6744a
Some cleaning up of the synthesis code.
by Jean-Marc Valin
· 11 years ago
ed01a59
Making exp_rotation1() use MAC16_16(), which saves a few cycles on ARM
by Jean-Marc Valin
· 11 years ago
ccec752
Silences unused parameter warning
by Jean-Marc Valin
· 11 years ago
ef0eac4
Moving the radix-2 to expose trivial twiddle factors
by Jean-Marc Valin
· 11 years ago
c8f62a4
Improving the accuracy of the fixed-point radix-3 and radix-5
by Jean-Marc Valin
· 11 years ago
e1f8462
Minor cleanup -- nothing to see here
by Jean-Marc Valin
· 11 years ago
05291fd
Fixed-point: slight accuracy improvement in the comb filter
by Jean-Marc Valin
· 11 years ago
30f52cb
Remove a SAVE_STACK that was pasted accidentally in the previous commit
by Jean-Marc Valin
· 11 years ago
e1dc1e2
Unifying scaling of fixed-point and float FFT
by Jean-Marc Valin
· 11 years ago
dbb96ab
Fixes C89 issue
by Jean-Marc Valin
· 11 years ago
4c1a90a
Getting rid of some negations
by Jean-Marc Valin
· 11 years ago
cc344fb
Slightly improving the accuracy of the fixed-point MDCT downscale
by Jean-Marc Valin
· 11 years ago
e0c00e2
Commit 99968ab was causing us to allocate too much stack in the MDCT
by Jean-Marc Valin
· 11 years ago
e43a0ab
Removes the separate 1/8N rotation in the (I)MDCT and unmerges the MDCT sizes
by Jean-Marc Valin
· 11 years ago
a5e3c8a
Inverse MDCT no longer requires any scratch space
by Jean-Marc Valin
· 11 years ago
e2bcb3f
Reverse the ordering of the FFT stage to optimize a degenerate radix-4 case.
by Jean-Marc Valin
· 11 years ago
c8f4e16
Merges the FFT scaling with the MDCT pre-rotate
by Jean-Marc Valin
· 11 years ago
153def2
Getting rid of the inverse FFT entirely
by Jean-Marc Valin
· 11 years ago
99968ab
Moving bitrev step to forward MDCT too
by Jean-Marc Valin
· 11 years ago
bc13bba
Applying the forward FFT gain up-front for fixed-point too
by Jean-Marc Valin
· 11 years ago
2e26b82
Moves the bitrev step to the IMDCT pre-rotation
by Jean-Marc Valin
· 11 years ago
306d7f5
fixed-point: slight (but free) accuracy improvement in compute_band_energies()
by Jean-Marc Valin
· 11 years ago
e0f2624
fixed-point: adds rounding to some shifts to eliminate bias
by Jean-Marc Valin
· 11 years ago
4a168eb
Remove useless code in alloc_trim_analysis()
by Jean-Marc Valin
· 11 years ago
5752d65
Minor fixed-point accuracy improvements that were completely free
by Jean-Marc Valin
· 11 years ago
91f8010
Removing indirections
by Jean-Marc Valin
· 11 years ago
5607d5d
Annotating pointer arguments with OPUS_RESTRICT and const
by Jean-Marc Valin
· 11 years ago
122971b
More NaN hardening in the analysis code
by Jean-Marc Valin
· 11 years ago
d5553e8
Using OPUS_COPY()/OPUS_CLEAR() in the decoder too
by Jean-Marc Valin
· 11 years ago
15edb78
Making NaN detection more robust to -ffast-math.
by Jean-Marc Valin
· 11 years ago
4fda6b0
Using celt_inner_prod() in compute_band_energies()
by Jean-Marc Valin
· 11 years ago
a9b7def
s/MAX16/MAX32/ in transient_analysis()
by Pedro Becerra
· 11 years ago
57cd849
Defining celt_inner_prod() and using it instead of explicit loops.
by Jean-Marc Valin
· 11 years ago
ff07200
Replaces inline copies and initialization with OPUS_*() macros.
by Jean-Marc Valin
· 11 years ago
0f869cb
Changes ABS16() and ABS32() to use fabs() in the float build
by Jean-Marc Valin
· 11 years ago
Next »