Gitiles
Code Review
Sign In
gerrit-public.fairphone.software
/
platform
/
external
/
libopus
/
320cf2e2cda1f9911a7625de1994cda28a52380d
/
libcelt
320cf2e
Re-organize spreading/folding constants.
by Timothy B. Terriberry
· 14 years ago
cd84e3d
Re-enabling post-filter on 2.5 ms frames
by Jean-Marc Valin
· 14 years ago
428a77d
More cleanups to compute_allocation().
by Timothy B. Terriberry
· 14 years ago
76ea41e
Give the bit we reserved to end skipping back when we don't use it.
by Timothy B. Terriberry
· 14 years ago
8e31ab3
Making lower sampling rates use the same allocation table
by Jean-Marc Valin
· 14 years ago
8e44767
Terminate the coding of skip bits at the last dynalloc boosted band. Otherwise the bitstream allows non-sensible behavior by the encoder (dynallocing bits into a band and then skipping it). This reduces skip bit overhead by about 2-3% at moderate bitrates with the current encoder.
by Gregory Maxwell
· 14 years ago
034172b
Making the encoder fail less often at stupidly low bit-rates
by Jean-Marc Valin
· 14 years ago
5c80391
Comments, low bit-rate busting avoidance
by Jean-Marc Valin
· 14 years ago
df6620e
Setting fine_priority for skipped bands
by Jean-Marc Valin
· 14 years ago
66c5ab4
Rebalance N=1 allocations during interp_bits2pulses().
by Timothy B. Terriberry
· 14 years ago
405e6a9
Change strategies for allocation hole prevention.
by Timothy B. Terriberry
· 14 years ago
7627b9f
More compute_allocation() fixes.
by Timothy B. Terriberry
· 14 years ago
4777f06
Store the total budget of compute_allocation in BITRES units.
by Timothy B. Terriberry
· 14 years ago
283a9b6
Don't stop force-skipping on the first manually skipped band.
by Timothy B. Terriberry
· 14 years ago
b2f5900
Move skip coding into interp_bits2pulses().
by Timothy B. Terriberry
· 14 years ago
7cbf168
More allocation function updates.
by Timothy B. Terriberry
· 14 years ago
ffe1057
Tim just rewrote half of the bit allocator -- hope it works now
by Jean-Marc Valin
· 14 years ago
9651ffd
This fixes a bunch of bit allocation bugs
by Jean-Marc Valin
· 14 years ago
fa50d3a
More negative bits prevention
by Jean-Marc Valin
· 14 years ago
dcacb73
Preventing negative bit allocation
by Jean-Marc Valin
· 14 years ago
6cbfbc3
Tuning for 2.5 ms frames
by Jean-Marc Valin
· 14 years ago
425cfd3
Fix 2-phase stereo in the encoder
by Jean-Marc Valin
· 14 years ago
70d30ff
Using overlap=shortMdctSize even for 2.5 ms frames.
by Jean-Marc Valin
· 14 years ago
546dfa1
Adapting the allocation trim based on the spectral tilt
by Jean-Marc Valin
· 14 years ago
dfd6e71
Adding some hysteresis on the folding threshold frequency
by Jean-Marc Valin
· 14 years ago
85bbab0
Fixes a fixed-point overflow in stereo angle calculation
by Jean-Marc Valin
· 14 years ago
fddc521
Completely new transient analysis algorithm
by Jean-Marc Valin
· 14 years ago
2b747c9
Simplifying the allocation code and making sure both channels get a bit
by Jean-Marc Valin
· 14 years ago
52dc66b
Take into accound the 32-bit limit in the VQ
by Jean-Marc Valin
· 14 years ago
79b34eb
Add API support for unconstrained VBR
by Jean-Marc Valin
· 14 years ago
9faf740
Keeping the allocation of the intensity-codec bands
by Jean-Marc Valin
· 14 years ago
a4badac
Making VBR slightly exceed the budget rather than fail
by Jean-Marc Valin
· 14 years ago
30165bb
Fixing the most obvious problems with the VBR code
by Jean-Marc Valin
· 14 years ago
e5e9aa7
Fixes some side-information rate control issues in VBR mode
by Jean-Marc Valin
· 14 years ago
e65978f
Adding a dual stereo option.
by Jean-Marc Valin
· 14 years ago
100ae8c
nothing to see here
by Jean-Marc Valin
· 14 years ago
1bfa18c
Fix totally broken bit allocation for non-mainstream modes (e.g. powers of two).
by Jean-Marc Valin
· 14 years ago
bad42a7
Better tuning for 5 ms frames
by Jean-Marc Valin
· 14 years ago
4b087df
Increasing resolution of the alloc trim
by Jean-Marc Valin
· 14 years ago
dd4bf63
Tuning higher bit-rate to behave a bit more like Vorbis.
by Jean-Marc Valin
· 14 years ago
4f177e8
Intensity stereo now in the bit-stream
by Jean-Marc Valin
· 14 years ago
fb03111
Allocate remaining bits with constant SNR offset
by Jean-Marc Valin
· 14 years ago
9b4ed5e
Implementing intensity stereo.
by Jean-Marc Valin
· 14 years ago
128d0c1
Tuning the allocation trim_offset for shorter frames
by Jean-Marc Valin
· 14 years ago
e6d7c08
Minor tuning to the last band
by Jean-Marc Valin
· 14 years ago
c992a0d
New bit allocation tuning now enabled by default
by Jean-Marc Valin
· 14 years ago
54d84c0
Bit allocation wotk.
by Jean-Marc Valin
· 14 years ago
0722c16
Good catch by John Ridges
by Jean-Marc Valin
· 14 years ago
ef2e650
Add coarse energy entropy model tuning.
by Timothy B. Terriberry
· 14 years ago
1ad93cf
Fixes several fixed-point overflows in the PLC code
by Jean-Marc Valin
· 14 years ago
e53c4bc
Fixes a silly fixed-point scaling PLC bug
by Jean-Marc Valin
· 14 years ago
d7231dd
Giving up on reusing the saved overlap in the PLC
by Jean-Marc Valin
· 14 years ago
bc4a002
PLC fixes
by Jean-Marc Valin
· 14 years ago
8f6d12b
Adding float|fixed suffixes to static mode files
by Jean-Marc Valin
· 14 years ago
6c12497
Increases the probability of alloc_trim==2 to reflect the latest changes
by Jean-Marc Valin
· 14 years ago
44a9600
Minor tuning
by Jean-Marc Valin
· 14 years ago
9743bf3
Switch iteration over channels to the do{}while(); construct in order to inform the compiler that the these loops execute at least once. (This results in more intelligent output from the clang static analysis tool and should also produce faster code on at least some architectures.)
by Gregory Maxwell
· 14 years ago
60c316b
Eliminate some promotions to double. A fair number of implicit promotions remain but they all involve math functions which exist only as double precision form in C89.
by Gregory Maxwell
· 14 years ago
fa1ca19
Make sure that we always code the sign of N=1 bands
by Jean-Marc Valin
· 14 years ago
a3a066c
Fixes some stereo issues where the right channel wasn't taken into account
by Jean-Marc Valin
· 14 years ago
35095c6
Squashed commit of the following:
by Jean-Marc Valin
· 14 years ago
bc2c454
Fixed the PLC, which had been broken for a while
by Jean-Marc Valin
· 14 years ago
fac6c98
Fix crashes with VBR for short duration frames and very low bitrates. This may, however, cause the encoder to violate the rate target at insanely low rates.
by Gregory Maxwell
· 14 years ago
a941147
Switch example tools to use VBR and 960 sample frames by default on the basis that if the user doesn't have any particular requirements that they probably want this.
by Gregory Maxwell
· 14 years ago
eedb422
Further simplications to compute_mdcts()
by Jean-Marc Valin
· 14 years ago
933dd83
De-interleaves the MDCT input and overlap memory.
by Jean-Marc Valin
· 14 years ago
9037757
Tuning the allocation trim
by Jean-Marc Valin
· 14 years ago
c40addc
Reworked the allocation trim to be absolute (in bits/sample) rather relative
by Jean-Marc Valin
· 14 years ago
fd54a99
Prevents a folding artefact due to "folding the folding"
by Jean-Marc Valin
· 14 years ago
9449165
Alternate tuning found using exp_tuning_knobs (disabled by default)
by Jean-Marc Valin
· 14 years ago
6bf15c5
Tuning the folding bands.
by Jean-Marc Valin
· 14 years ago
0110301
allocation trim doesn't make sense for stereo after all
by Jean-Marc Valin
· 14 years ago
c5792de
First shot at automatically adjusting the "allocation trim" for stereo.
by Jean-Marc Valin
· 14 years ago
5790fba
Simplifying transient_analysis() now that we don't care about the time window
by Jean-Marc Valin
· 14 years ago
7fff572
Making sure we can use up to 128 pulses.
by Jean-Marc Valin
· 14 years ago
d50b628
Removing renormalise_bands()
by Jean-Marc Valin
· 14 years ago
27fc0c4
s/const/static const/
by Jean-Marc Valin
· 14 years ago
7a08ddd
Removing both the transient window and the mdct_weight_shift.
by Jean-Marc Valin
· 14 years ago
35fceef
Turning off time-domain window pending decision on what to do with it.
by Jean-Marc Valin
· 14 years ago
e0aa9d1
Removing dead code found by LLVM's static analysis
by Jean-Marc Valin
· 14 years ago
e0ae980
Optimised float->int cast for Win64
by John Ridges
· 14 years ago
bb918ef
Fixed fixed-point overflow for 2.5 ms frames.
by Jean-Marc Valin
· 14 years ago
3295b5d
Folding code moved to quant_band() to prevent duplication.
by Jean-Marc Valin
· 14 years ago
c4711e4
More VQ search simplifications
by Jean-Marc Valin
· 14 years ago
bbc010c
Removing yshift from the fixed-point VQ search
by Jean-Marc Valin
· 14 years ago
3d25e71
More simplifications to the VQ search
by Jean-Marc Valin
· 14 years ago
aca7a63
VQ search no longer needs to put more than one pulse at a time
by Jean-Marc Valin
· 14 years ago
dfb3e68
tweaks to the split delta for transients
by Jean-Marc Valin
· 14 years ago
4d2d9fc
Transient detection fix for the case where a transient occurs during the overlap
by Jean-Marc Valin
· 14 years ago
54fb7e5
Tuning the bias for the L1 tf metric
by Jean-Marc Valin
· 14 years ago
8852563
Unb0rked the VBR code: several units issues that made VBR completely useless
by Jean-Marc Valin
· 14 years ago
8200b2d
Modifying the L1 metric for tf analysis by merging together all frequencies.
by Jean-Marc Valin
· 14 years ago
4a8c1f1
Making it possible to have two transients in a row for longer frames
by Jean-Marc Valin
· 14 years ago
2794b63
Adding a high-pass filter in front of the transient analysis code
by Jean-Marc Valin
· 14 years ago
ccd51fe
No longer filling the stereo side with noise.
by Jean-Marc Valin
· 14 years ago
ccd5a61
Taking into account TF decisions for VBR
by Jean-Marc Valin
· 14 years ago
0f8fc0b
New TF decision code based on L1-norm. Needs more work.
by Jean-Marc Valin
· 14 years ago
844f00d
Don't take into account silence for the coarse energy means
by Jean-Marc Valin
· 14 years ago
a8160dd
Remove lower bound on bit-rate for tf signalling
by Jean-Marc Valin
· 14 years ago
f1fea66
Dynamic allocation before VBR
by Jean-Marc Valin
· 14 years ago
Next »