Performance improvements:
quant_bands:
with: 681 decicycles in quant_bands, 8388453 runs, 155 skips
without: 1190 decicycles in quant_bands, 8388386 runs, 222 skips
Around 42% for the function
Twoloop coder:
abs_pow34:
with/without: 7.82s/8.17s
Around 4% for the entire encoder
Both:
with/without: 7.15s/8.17s
Around 12% for the entire encoder
Fast coder:
abs_pow34:
with/without: 3.40s/3.77s
Around 10% for the entire encoder
Both:
with/without: 3.02s/3.77s
Around 20% faster for the entire encoder
Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
Tested-by: Michael Niedermayer <michael@niedermayer.cc>
Reviewed-by: James Almer <jamrial@gmail.com>
Using lfg was an overkill in this case where the random numbers
were only used for encoder descisions. Should increase result
uniformity between different FPUs and gives a slight speedup.
Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
The idea is to use ffmath.h for internal implementations of math functions.
Currently, it is used for variants of libm functions, but is by no means
limited to such things.
Note that this is not exported; use lavu/mathematics for such purposes.
Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanag@gmail.com>
This is ~2x faster for y not an integer on Haswell+GCC, and should
generally be faster due to the fact that anyway powf essentially does
this under the hood. Made an inline function in lavu/internal.h for this
purpose.
Note that there are some accuracy differences, that should generally be
negligible. In particular, FATE still passes on this platform.
Results in ~ 7% speedup in aac encoding with -march=native, Haswell+GCC.
before:
ffmpeg -i sin.flac -acodec aac -y sin_new.aac 6.05s user 0.06s system 104% cpu 5.821 total
after:
ffmpeg -i sin.flac -acodec aac -y sin_new.aac 5.67s user 0.03s system 105% cpu 5.416 total
This is also faster than an alternative approach that pulls in powf, gets rid of
the crufty NaN checks and other special cases, exploits knowledge about the intervals, etc.
This of course does not exclude smarter approaches; just suggests that
there would need to be significant work on this front of lower utility than
searches for hotspots elsewhere.
Reviewed-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanag@gmail.com>
This ensures gcc does not create unnecessary
loads or stores and possibly even does not vectorize
the negation.
Speeds up mp3 to aac transcoding with default settings
by 10% when using "gcc (Debian 5.3.1-10) 5.3.1 20160224".
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
I cannot see any point whatsoever to use
double here instead of float, the results
are likely identical in all cases..
Using float allows for much more
efficient use of SIMD.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
It makes no sense whatsoever to do this at each function call; we
already have a table for this.
Yields a 2x improvement in find_min_book (x86-64, Haswell+GCC):
ffmpeg -i sin.flac -acodec aac -y sin.aac
find_min_book
old
605 decicycles in find_min_book, 8388453 runs, 155 skips.9x
606 decicycles in find_min_book,16776912 runs, 304 skips.9x
607 decicycles in find_min_book,33553819 runs, 613 skips.2x
607 decicycles in find_min_book,67107668 runs, 1196 skips.3x
607 decicycles in find_min_book,134215360 runs, 2368 skips3x
new
359 decicycles in find_min_book, 8388552 runs, 56 skips.3x
360 decicycles in find_min_book,16777112 runs, 104 skips.1x
361 decicycles in find_min_book,33554218 runs, 214 skips.4x
361 decicycles in find_min_book,67108381 runs, 483 skips.5x
361 decicycles in find_min_book,134216725 runs, 1003 skips5x
and more importantly a non-negligible speedup (~ 8%) to overall AAC encoding:
old:
ffmpeg -i sin.flac -acodec aac -strict -2 -y sin_new.aac 6.82s user 0.03s system 104% cpu 6.565 total
new:
ffmpeg -i sin.flac -acodec aac -strict -2 -y sin_old.aac 6.24s user 0.03s system 104% cpu 5.993 total
This also improves accuracy of the expression by ~ 2 ulp in some cases.
Reviewed-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Reviewed-by: Rostislav Pehlivanov <atomnuker@gmail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanag@gmail.com>
This is quite an accurate approximation; testing shows ~ 2ulp error in
the floating point result. Tested with FATE.
Alternatively, if one wants "full accuracy", one can use powf, or sqrt
instead of sqrtf. With powf, one gets 1 ulp error (theoretically should be 0, as
0.75 is exactly representable) on GNU libm, with sqrt, 0 ulp error.
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Abstracted into pos_pow34 utility function
Signed-off-by: Claudio Freire <klaussfreire@gmail.com>
PSNR doesn't change as expected. The AAC spec doesn't really say
anything about how exactly to generate noise.
Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
This patch does 4 things, all of which interact and thus it
woudln't be possible to commit them separately without causing
either quality regressions or assertion failures.
Fate comparison targets don't all reflect improvements in
quality, yet listening tests show substantially improved quality
and stability.
1. Increase SF range utilization.
The spec requires SF delta values to be constrained within the
range -60..60. The previous code was applying that range to
the whole SF array and not only the deltas of consecutive values,
because doing so requires smarter code: zeroing or otherwise
skipping a band may invalidate lots of SF choices.
This patch implements that logic to allow the coders to utilize
the full dynamic range of scalefactors, increasing quality quite
considerably, and fixing delta-SF-related assertion failures,
since now the limitation is enforced rather than asserted.
2. PNS tweaks
The previous modification makes big improvements in twoloop's
efficiency, and every time that happens PNS logic needs to be
tweaked accordingly to avoid it from stepping all over twoloop's
decisions. This patch includes modifications of the sort.
3. Account for lowpass cutoff during PSY analysis
The closer PSY's allocation is to final allocation the better
the quality is, and given these modifications, twoloop is now
very efficient at avoiding holes. Thus, to compute accurate
thresholds, PSY needs to account for the lowpass applied
implicitly during twoloop (by zeroing high bands).
This patch makes twoloop set the cutoff in psymodel's context
the first time it runs, and makes PSY account for it during
threshold computation, making PE and threshold computations
closer to the final allocation and thus achieving better
subjective quality.
4. Tweaks to RC lambda tracking loop in relation to PNS
Without this tweak some corner cases cause quality regressions.
Basically, lambda needs to react faster to overall bitrate
efficiency changes since now PNS can be quite successful in
enforcing maximum bitrates, when PSY allocates too many bits
to the lower bands, suppressing the signals RC logic uses to
lower lambda in those cases and causing aggressive PNS.
This tweak makes PNS much less aggressive, though it can still
use some further tweaks.
Also update MIPS specializations and adjust fuzz
Also in lavc/mips/aacpsy_mips.h: remove trailing whitespace
It didn't work out because of the exceptions that needed to be made
for the "-1" cases and was overall more confusing that just manually
checking and setting options for each profile.
This commit adds the ability for a profile to set the default
options, as well as for the user to override such options
by simply stating them in the command line while still keeping
the same profile, as long as those options are still permitted by
the profile.
Example: setting the profile to aac_low (the default) will turn
PNS and IS on. They can be disabled by -aac_pns 0 and -aac_is 0,
respectively. Turning on -aac_pred 1 will cause the profile to be
elevated to aac_main, as long as no options forbidding aac_main
have been entered (like AAC-LTP, which will be pushed soon).
A useful feature is that by setting the profile to mpeg2_aac_low,
all MPEG4 features will be disabled and if the user tries to enable
them then the program will exit with an error. This profile is
signalled with the same bitstream as aac_low (MPEG4) but some devices
and decoders will fail if any MPEG4 features have been enabled.
This finalizes merging of the work in the patches in ticket #2686.
Improvements to twoloop and RC logic are extensive.
The non-exhaustive list of twoloop improvments includes:
- Tweaks to distortion limits on the RD optimization phase of twoloop
- Deeper search in twoloop
- PNS information marking to let twoloop decide when to use it
(turned out having the decision made separately wasn't working)
- Tonal band detection and priorization
- Better band energy conservation rules
- Strict hole avoidance
For rate control:
- Use psymodel's bit allocation to allow proper use of the bit
reservoir. Don't work against the bit reservoir by moving lambda
in the opposite direction when psymodel decides to allocate more/less
bits to a frame.
- Retry the encode if the effective rate lies outside a reasonable
margin of psymodel's allocation or the selected ABR.
- Log average lambda at the end. Useful info for everyone, but especially
for tuning of the various encoder constants that relate to lambda
feedback.
Psy:
- Do not apply lowpass with a FIR filter, instead just let the coder
zero bands above the cutoff. The FIR filter induces group delay,
and while zeroing bands causes ripple, it's lost in the quantization
noise.
- Experimental VBR bit allocation code
- Tweak automatic lowpass filter threshold to maximize audio bandwidth
at all bitrates while still providing acceptable, stable quality.
I/S:
- Phase decision fixes. Unrelated to #2686, but the bugs only surfaced
when the merge was finalized. Measure I/S band energy accounting for
phase, and prevent I/S and M/S from being applied both.
PNS:
- Avoid marking short bands with PNS when they're part of a window
group in which there's a large variation of energy from one window
to the next. PNS can't preserve those and the effect is extremely
noticeable.
M/S:
- Implement BMLD protection similar to the specified in
ISO-IEC/13818:7-2003, Appendix C Section 6.1. Since M/S decision
doesn't conform to section 6.1, a different method had to be
implemented, but should provide equivalent protection.
- Move the decision logic closer to the method specified in
ISO-IEC/13818:7-2003, Appendix C Section 6.1. Specifically,
make sure M/S needs less bits than dual stereo.
- Don't apply M/S in bands that are using I/S
Now, this of course needed adjustments in the compare targets and
fuzz factors of the AAC encoder's fate tests, but if wondering why
the targets go up (more distortion), consider the previous coder
was using too many bits on LF content (far more than required by
psy), and thus those signals will now be more distorted, not less.
The extra distortion isn't audible though, I carried extensive
ABX testing to make sure.
A very similar patch was also extensively tested by Kamendo2 in
the context of #2686.
As well as tables littered everywhere, functions were spread
out all across the encoder's files. This moves them to a single
place where they can be used by either the encoder's main files
or additional encoder files. Additionally, it changes the type
of some to 'inline' to enable us to simply put them in a header
file and possibly gain some speed due to compiler optimizations.
Signed-off-by: Claudio Freire <klaussfreire@gmail.com>
Rostislav Pehlivanov
Ganesh Ajjanagadde
Reimar Döffinger
Claudio Freire
Ganesh Ajjanagadde