Instead use float one by default for sample rate conversions.
The s16p internal transfer format produces visible and hearable
quantization artifacts.
Signed-off-by: Paul B Mahol <onemda@gmail.com>
for S8 we continue to use S16 as it should have enough precision
Fate is adjusted so bitexactness is maintained between mips/arm/x86
if more tests became bit-inexact on some platform, the same change
can be done to them
The use of higher precision and float intermediates inevitably
leads to more differences between platforms.
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
We've been fuzzing torchvision with [sydr-fuzz](https://github.com/ispras/oss-sydr-fuzz)
and found out of bounds error in ffmpeg project at audioconvert.c:151.
To prevent error we need to fix checks for in and out fmt in swr_init.
Signed-off-by: Eli Kobrin <kobrineli@ispras.ru>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
If it's unsupported or invalid, then there's no point trying to rebuild it
using a value that may have been derived from the same layout to begin with.
Move the checks before the attempts at copying the layout while at it.
Fixes ticket #9908.
Signed-off-by: James Almer <jamrial@gmail.com>
This avoids having to rebuild big files every time FFMPEG_VERSION
changes (which it does with every commit).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
If memory allocation fails, ERROR(ENOMEM) '-12' will be returned.
When resample() is done first, the negative size param would cause buffer-overflow and SEGV in swri_rematrix().
When swri_rematrix() is run first, resample() would not cause an error but Err num as a wrong parameter passing.
Err num should be returned immediately. And remove an unneeded term from an assert.
coredump info:
#0 0x499517 in posix_memalign (/home/r1/ffmpeg/ffmpeg_4.4.1+0x499517)
#1 0x6c1f0b4 in av_malloc /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavutil/mem.c:86:9
#2 0x6c208fe in av_mallocz /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavutil/mem.c:239:17
#3 0x6c207ad in av_mallocz_array /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavutil/mem.c:195:12
#4 0x654b2e5 in swri_realloc_audio /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libswresample/swresample.c:418:14
#5 0x654f9a1 in swr_convert_internal /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libswresample/swresample.c:601:17
#6 0x654d2c0 in swr_convert /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libswresample/swresample.c:766:19
#7 0x186cf56 in flush_frame /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/af_aresample.c:251:13
#8 0x186a454 in request_frame /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/af_aresample.c:288:20
#9 0x787d9c in ff_request_frame_to_filter /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/avfilter.c:459:15
#10 0x7877f1 in forward_status_change /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/avfilter.c:1257:19
#11 0x77ed7e in ff_filter_activate_default /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/avfilter.c:1288:20
#12 0x77e4e1 in ff_filter_activate /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/avfilter.c:1441:11
#13 0x793b3f in ff_filter_graph_run_once /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/avfiltergraph.c:1403:12
#14 0x7a7bee in get_frame_internal /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/buffersink.c:131:19
#15 0x7a7287 in av_buffersink_get_frame_flags /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/buffersink.c:142:12
#16 0x792888 in avfilter_graph_request_oldest /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/libavfilter/avfiltergraph.c:1356:17
#17 0x5d07df in transcode_from_filter /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/fftools/ffmpeg.c:4639:11
#18 0x59e557 in transcode_step /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/fftools/ffmpeg.c:4729:20
#19 0x593970 in transcode /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/fftools/ffmpeg.c:4805:15
#20 0x58f7a4 in main /home/r1/ffmpeg/ffmpeg-4.4.1/build/src/fftools/ffmpeg.c:5010:9
#21 0x7f6fd2dee0b2 in __libc_start_main /build/glibc-eX1tMB/glibc-2.31/csu/../csu/libc-start.c:308:16
SUMMARY: AddressSanitizer: negative-size-param (/home/r1/ffmpeg/ffmpeg_4.4.1+0x497e67) in __asan_memcpy
Reported-by: TOTE Robot <oslab@tsinghua.edu.cn>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
SWR_CH_MAX is internal only and the arrays are therefore not required
to have that many elements (and they typically don't do it). So remove
this potentially confusing hint.
(Newer versions of GCC emit -Warray-parameter= warnings for this,
because the definition with explicit size differs from the declaration
(which leaves the size unspecified); this is IMO a false-positive,
because definition and declaration didn't conflict, but anyway it is
fixed by this commit.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Removed +len1 in call to s->mix_2_1_f() as I found no logical explanation for it. After removal, problem was gone.
Signed-off-by: Hendrik Schreiber <hs@tagtraum.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
use fltp when doing s32 -> s32 resampling
because s32p has no simd optimization
benchmark:
old 17.913s
new 7.584s (use fma3)
Reviewed-by: wm4 <nfxjfg@googlemail.com>
Signed-off-by: Muhammad Faiz <mfcc64@gmail.com>
so tsf option in aresample will have effect
previously tsf/internal_sample_format had no effect
fate is updated
s32p previously used fltp internally
dblp previously used fltp/dblp internally
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Muhammad Faiz <mfcc64@gmail.com>
Fixes undefined operation
Fixes part of 668007-media
Found-by: Matt Wolenetz <wolenetz@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
give high quality resampling
as good as with linear_interp=on
as fast as without linear_interp=on
tested visually with ffplay
ffplay -f lavfi "aevalsrc='sin(10000*t*t)', aresample=osr=48000, showcqt=gamma=5"
ffplay -f lavfi "aevalsrc='sin(10000*t*t)', aresample=osr=48000:linear_interp=on, showcqt=gamma=5"
ffplay -f lavfi "aevalsrc='sin(10000*t*t)', aresample=osr=48000:exact_rational=on, showcqt=gamma=5"
slightly speed improvement
for fair comparison with -cpuflags 0
audio.wav is ~ 1 hour 44100 stereo 16bit wav file
ffmpeg -i audio.wav -af aresample=osr=48000 -f null -
old new
real 13.498s 13.121s
user 13.364s 12.987s
sys 0.131s 0.129s
linear_interp=on
old new
real 23.035s 23.050s
user 22.907s 22.917s
sys 0.119s 0.125s
exact_rational=on
real 12.418s
user 12.298s
sys 0.114s
possibility to decrease memory usage if soft compensation is ignored
Signed-off-by: Muhammad Faiz <mfcc64@gmail.com>
or if no rematrix and no resampling is performed and the input is 16bit
note reampling and rematrix itself always use more than 16bit internally
the "internal" sampling format is the format between these steps
Its unlikely the difference from this commit is audible in any case
unless there is some bug either before or after the change.
but multiple people prefer this and it slightly improves the precission
of computations.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This avoids leaks if the user doest call swr_close() after a failed init
Found-by: James Almer <jamrial@gmail.com>
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Previous version reviewed-by: Pavel Koshevoy <pkoshevoy@gmail.com>
Previous version reviewed-by: wm4 <nfxjfg@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Fix crash when doing 8 ch conversion from apps compiled with MSVS
Thanks to Ronald for giving this hint:
https://ffmpeg.org/pipermail/ffmpeg-devel/2015-May/173049.html
Reviewed-by: "Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Paul B Mahol
Eli Kobrin
James Almer
rcombs
Andreas Rheinhardt
Martin Storsjö
Yu Yang
Carl Eugen Hoyos
Michael Niedermayer
Hendrik Schreiber
Muhammad Faiz
Michael Niedermayer
Clément Bœsch
Ganesh Ajjanagadde
Rainer Hochecker
Timothy Gu