ffmpeg.git
4 years agotools/bookmarklets: HTML code fixes
Reto Kromer [Thu, 24 Dec 2015 14:00:46 +0000 (15:00 +0100)]
tools/bookmarklets: HTML code fixes

Minor HTML5 code changes.

Closes #167

4 years agoavfilter/af_anequalizer: make cliping filter type actually useful
Paul B Mahol [Fri, 25 Dec 2015 14:03:46 +0000 (15:03 +0100)]
avfilter/af_anequalizer: make cliping filter type actually useful

Previously result was ignored.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavfilter/af_anequalizer: use pow instead of exp10
Paul B Mahol [Fri, 25 Dec 2015 14:00:06 +0000 (15:00 +0100)]
avfilter/af_anequalizer: use pow instead of exp10

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavfilter: add high-order parametric multiband equalizer filter
Paul B Mahol [Sun, 20 Dec 2015 08:50:35 +0000 (09:50 +0100)]
avfilter: add high-order parametric multiband equalizer filter

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agodoc/muxers/mpegts: update doc after adding flag to embed an AC-3/E-AC-3 ES the DVB way
Stefan Pöschel [Fri, 25 Dec 2015 09:33:35 +0000 (10:33 +0100)]
doc/muxers/mpegts: update doc after adding flag to embed an AC-3/E-AC-3 ES the DVB way

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoavcodec: add OpenJPEG 2.x compatibility
Michael Bradshaw [Mon, 2 Nov 2015 03:11:12 +0000 (19:11 -0800)]
avcodec: add OpenJPEG 2.x compatibility

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoffplay: Fix auto insertion point of rotation filter
wanzhang [Wed, 29 Jul 2015 17:16:43 +0000 (01:16 +0800)]
ffplay: Fix auto insertion point of rotation filter

Fixes watermark rotation for videos using auto rotation

Closes #141

4 years agoswscale/utils: Fix intermediate format for cascaded alpha downscaling
Michael Niedermayer [Thu, 24 Dec 2015 20:46:15 +0000 (21:46 +0100)]
swscale/utils: Fix intermediate format for cascaded alpha downscaling

Fixes Ticket4926

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoavformat/mpegts: consider stream_type 4 just a hint toward mp3 and not definite
Michael Niedermayer [Thu, 24 Dec 2015 19:32:11 +0000 (20:32 +0100)]
avformat/mpegts: consider stream_type 4 just a hint toward mp3 and not definite

Fixes Ticket 4864

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoavcodec/ac3dec: Print the value of out of range exponents
Michael Niedermayer [Thu, 24 Dec 2015 16:46:12 +0000 (17:46 +0100)]
avcodec/ac3dec: Print the value of out of range exponents

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoswr/resample: use av_clip_int16 instead of av_clip
Ganesh Ajjanagadde [Wed, 23 Dec 2015 22:48:40 +0000 (14:48 -0800)]
swr/resample: use av_clip_int16 instead of av_clip

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agoavfilter/af_sofalizer: make virtual speaker positioning supports all channel layouts
Paul B Mahol [Thu, 24 Dec 2015 16:53:11 +0000 (17:53 +0100)]
avfilter/af_sofalizer: make virtual speaker positioning supports all channel layouts

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agox86/vf_blend: add sse2 versions of blend_difference and blend_negation
James Almer [Thu, 24 Dec 2015 02:55:56 +0000 (23:55 -0300)]
x86/vf_blend: add sse2 versions of blend_difference and blend_negation

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
4 years agox86/vf_blend: make all functions work on x86_32
James Almer [Thu, 24 Dec 2015 02:54:33 +0000 (23:54 -0300)]
x86/vf_blend: make all functions work on x86_32

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
4 years agox86/vf_blend: simplify using macros
James Almer [Thu, 24 Dec 2015 02:51:45 +0000 (23:51 -0300)]
x86/vf_blend: simplify using macros

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
4 years agox86/vf_maskedmerge: make ff_maskedmerge8_sse2 work on x86_32
James Almer [Wed, 23 Dec 2015 23:37:37 +0000 (20:37 -0300)]
x86/vf_maskedmerge: make ff_maskedmerge8_sse2 work on x86_32

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
4 years agoavformat/mpegtsenc: add flag to embed an AC-3/E-AC-3 ES the DVB way
Stefan Pöschel [Sun, 13 Dec 2015 10:54:32 +0000 (11:54 +0100)]
avformat/mpegtsenc: add flag to embed an AC-3/E-AC-3 ES the DVB way

So far an AC-3 elementary stream is refered to in the PMT according to
System A (ATSC). An E-AC-3 ES in contrast is embedded the System B (DVB) way.
To fix this inconsistency, this commit changes the default E-AC-3 behaviour to
use the ATSC way, too. Furthermore a new flag is added to optionally select the
DVB way (regarding both codecs and possible further differences in the future).

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoavfilter/x86/vf_maskedmerge: Clear upper part of width
Michael Niedermayer [Wed, 23 Dec 2015 20:57:31 +0000 (21:57 +0100)]
avfilter/x86/vf_maskedmerge: Clear upper part of width

Fixes crash
Fixes: Ticket5055

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoavfilter/af_biquads: display clipping warnings once per filtered frame
Paul B Mahol [Wed, 23 Dec 2015 18:15:25 +0000 (19:15 +0100)]
avfilter/af_biquads: display clipping warnings once per filtered frame

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agolavu/libm: add exp10 support
Ganesh Ajjanagadde [Tue, 22 Dec 2015 03:05:00 +0000 (19:05 -0800)]
lavu/libm: add exp10 support

exp10 is a function available in GNU libm. Looks like no other common
libm has it. This adds support for it to FFmpeg.

There are essentially 2 ways of handling the fallback:
1. Using pow(10, x)
2. Using exp2(M_LOG2_10 * x).

First one represents a Pareto improvement, with no speed or accuracy
regression anywhere, but speed improvement limited to GNU libm.

Second one represents a slight accuracy loss (relative error ~ 1e-13)
for non GNU libm. Speedup of > 2x is obtained on non GNU libm platforms,
~30% on GNU libm. These are "average case numbers", another benefit is
the lack of triggering of the well-known terrible worst case paths
through pow.

Based on reviews, second one chosen. Comment added accordingly.

Reviewed-by: Hendrik Leppkes <h.leppkes@gmail.com>
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/af_aemphasis: remove unnecessary complex number usage
Ganesh Ajjanagadde [Tue, 22 Dec 2015 01:12:04 +0000 (17:12 -0800)]
lavfi/af_aemphasis: remove unnecessary complex number usage

complex is not available on all platforms. Furthermore, it is trivial to
rewrite complex number expressions to real arithmetic, and in fact
sometimes advantageous for performance reasons: by wrapping as a complex,
one forces a particular Cartesian representation that is not necessarily optimal for the purpose.

Configure dependencies also removed, and aemphasis is now available across
all platforms.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agoavformat/subtitles: treat negative duration like unknown duration
wm4 [Wed, 23 Dec 2015 14:08:58 +0000 (15:08 +0100)]
avformat/subtitles: treat negative duration like unknown duration

Fixes a specific srt sample, which has an event with negative duration.
libavcodec will convert an event with negative duration to an ASS event
which will be displayed forever, which is not wanted here.

Treat negative duration always as unknown duration instead, and show it
until the next subtitle event.

4 years agoavcodec/s302menc: comment out allowed channel layouts
Paul B Mahol [Wed, 23 Dec 2015 11:46:23 +0000 (12:46 +0100)]
avcodec/s302menc: comment out allowed channel layouts

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavcodec/libschroedingerenc: add missing AVClass to private context
Hendrik Leppkes [Wed, 23 Dec 2015 09:22:00 +0000 (10:22 +0100)]
avcodec/libschroedingerenc: add missing AVClass to private context

Fixes ticket #5104.

4 years agodiracdec: add missing check for pixel_range_index
Andreas Cadhalpun [Tue, 22 Dec 2015 18:44:00 +0000 (19:44 +0100)]
diracdec: add missing check for pixel_range_index

This fixes an out-of-bounds read introduced in commit 0379603.

Reviewed-by: Kieran Kunhya <kierank@obe.tv>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agomlvdec: validate bits_per_coded_sample
Andreas Cadhalpun [Sat, 19 Dec 2015 22:45:00 +0000 (23:45 +0100)]
mlvdec: validate bits_per_coded_sample

A negative bits_per_coded_sample doesn't make sense.
If it is too large, the size calculation for av_get_packet overflows,
resulting in allocation of a too small buffer.

Also make sure width and height are sane.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agoavformat/movenc-test: Make format static
Michael Niedermayer [Tue, 22 Dec 2015 16:12:02 +0000 (17:12 +0100)]
avformat/movenc-test: Make format static

Makes greping for non static consts easier

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoavformat/mov: Update handbrake_version threshold for full mp3 parsing
Michael Niedermayer [Tue, 22 Dec 2015 15:19:44 +0000 (16:19 +0100)]
avformat/mov: Update handbrake_version threshold for full mp3 parsing

Fixes: Endangered\ Species\ 1x01\ Collecting\ Merl.mp4

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agolavfi: make request_frame() non-recursive.
Nicolas George [Wed, 30 Sep 2015 14:11:57 +0000 (16:11 +0200)]
lavfi: make request_frame() non-recursive.

Instead of calling the input filter request_frame() method,
ff_request_frame() now marks the link and returns immediately.
buffersink is changed to activate the marked filters until
a frame is obtained.

4 years agolavfi: replace link.closed by link.status.
Nicolas George [Thu, 24 Sep 2015 08:07:42 +0000 (10:07 +0200)]
lavfi: replace link.closed by link.status.

The status field can carry any error code instead of just EOF.
Also only update it through a wrapper function and provide a timestamp.
Update the few filters that used it directly.

4 years agolavfi: deprecate avfilter_link_set_closed().
Nicolas George [Thu, 24 Sep 2015 08:06:26 +0000 (10:06 +0200)]
lavfi: deprecate avfilter_link_set_closed().

Applications are not supposed to mess with links,
they should close the sinks.
Furthermore, this function does not distinguish what end
of the link caused the close and does not have a timestamp.

4 years agolavfi: add link.current_pts field.
Nicolas George [Thu, 24 Sep 2015 08:05:42 +0000 (10:05 +0200)]
lavfi: add link.current_pts field.

4 years agolavfi: rename link.current_pts to current_pts_us.
Nicolas George [Thu, 24 Sep 2015 07:30:05 +0000 (09:30 +0200)]
lavfi: rename link.current_pts to current_pts_us.

This field is used for fast comparison between link ages,
it is in AV_TIME_BASE units, in other words microseconds,
µs =~ us.
Renaming it allows a second field in link time base units.

4 years agolavfi/vf_mpdecimate: remove request_frame().
Nicolas George [Sun, 29 Nov 2015 12:06:28 +0000 (13:06 +0100)]
lavfi/vf_mpdecimate: remove request_frame().

It is no longer needed since looping is not necessary.

4 years agoavcodec/ass: check for av_mallocz() failure
Michael Niedermayer [Tue, 22 Dec 2015 13:41:27 +0000 (14:41 +0100)]
avcodec/ass: check for av_mallocz() failure

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoMAINTAINERS: add Eran Kornblau for aes_ctr and movenccenc
erankor [Tue, 22 Dec 2015 07:20:34 +0000 (09:20 +0200)]
MAINTAINERS: add Eran Kornblau for aes_ctr and movenccenc

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agovaapi: Add VP9 hwaccell support
Timo Rothenpieler [Sat, 19 Dec 2015 19:31:00 +0000 (20:31 +0100)]
vaapi: Add VP9 hwaccell support

Signed-off-by: Timo Rothenpieler <timo@rothenpieler.org>
4 years agolavf/rmdec: Use correct format specifier for int64_t.
Carl Eugen Hoyos [Tue, 22 Dec 2015 08:29:08 +0000 (09:29 +0100)]
lavf/rmdec: Use correct format specifier for int64_t.

Fixes ticket #5100.

4 years agoAAC encoder: fix possible assertion failure in PNS
Claudio Freire [Tue, 22 Dec 2015 08:26:12 +0000 (05:26 -0300)]
AAC encoder: fix possible assertion failure in PNS

Fix possible SF delta violation that would cause an
eventual assertion failure in some corner cases (esp
on very low bitrates) when marking bands for PNS due
to misuse of the sf_delta utilities

4 years agoaacenc_is: rename variable
Rostislav Pehlivanov [Mon, 21 Dec 2015 17:23:38 +0000 (17:23 +0000)]
aacenc_is: rename variable

'erf' is far from the best name for a variable and is not very
descriptive since the actual variable points to the comparitively best
IS phase. Therefore rename it to 'best'.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
4 years agolavu/libm: add erf hack and make dynaudnorm available everywhere
Ganesh Ajjanagadde [Sun, 20 Dec 2015 02:17:03 +0000 (18:17 -0800)]
lavu/libm: add erf hack and make dynaudnorm available everywhere

Source code is from Boost:
http://www.boost.org/doc/libs/1_46_1/boost/math/special_functions/erf.hpp
with appropriate modifications for FFmpeg.

Tested on interval -6 to 6 (beyond which it saturates), +/-NAN, +/-INFINITY
under -fsanitize=undefined on clang to test for possible undefined behavior.

This function turns out to actually be essentially as accurate and faster than the
libm (GNU/BSD's/Mac OS X), and I can think of 3 reasons why upstream
does not use this:
1. They are not aware of it.
2. They are concerned about licensing - this applies especially to GNU
libm.
3. They do not know and/or appreciate the benefits of rational
approximations over polynomial approximations. Boost uses them to great
effect, see e.g swr/resample for bessel derived from them, which is also
similarly superior to libm variants.

First, performance.
sample benchmark (clang -O3, Haswell, GNU/Linux):

3e8 values evenly spaced from 0 to 6
time (libm):
./test  13.39s user 0.00s system 100% cpu 13.376 total
time (boost based):
./test  9.20s user 0.00s system 100% cpu 9.190 total

Second, accuracy.
1e8 eval pts from 0 to 6
maxdiff (absolute): 2.2204460492503131e-16
occuring at point where libm erf is correctly rounded, this is not.

Illustration of superior rounding of this function:
arg   : 0.83999999999999997
erf   : 0.76514271145499457
boost : 0.76514271145499446
real  : 0.76514271145499446

i.e libm is actually incorrectly rounded. Note that this is clear from:
https://github.com/JuliaLang/openlibm/blob/master/src/s_erf.c (the Sun
implementation used by both BSD and GNU libm's), where only 1 ulp is
guaranteed.

Reasons it is not easy/worthwhile to create a "correctly rounded"
variant of this function (i.e 0.5ulp):
1. Upstream libm's don't do it anyway, so we can't guarantee this unless
we force this implementation on all platforms. This is not easy, as the
linker would complain unless measures are taken.
2. Nothing in FFmpeg cares or can care about such things, due to the
above and FFmpeg's nature.
3. Creating a correctly rounded function will in practice need some use of long
double/fma. long double, although C89/C90, unfortunately has problems on
ppc. This needs fixing of toolchain flags/configure. In any case this
will be slower for miniscule gain.

Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavf/avformat: add av_warn_unused_result to avformat_write_header
Ganesh Ajjanagadde [Mon, 21 Dec 2015 16:36:46 +0000 (08:36 -0800)]
lavf/avformat: add av_warn_unused_result to avformat_write_header

May be useful as a defense, see e.g c62d1780fff8a1997dd1707bbc557efc8fe41e3c.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavd/pulse_audio_enc: replace lround by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:59:12 +0000 (14:59 -0500)]
lavd/pulse_audio_enc: replace lround by lrint

Here it is mostly a cosmetic change, but there might be benefits in that
there are no compat hacks for lround, while there are for lrint.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/vf_histogram: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:56:18 +0000 (14:56 -0500)]
lavfi/vf_histogram: replace round by lrint

lrint is at least as fast, uses a superior rounding mode, and avoids an
implicit cast.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/af_dynaudnorm: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:51:09 +0000 (14:51 -0500)]
lavfi/af_dynaudnorm: replace round by lrint

lrint is at least as fast, uses a superior rounding mode, and avoids an
implicit cast.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/vf_crop: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:50:00 +0000 (14:50 -0500)]
lavfi/vf_crop: replace round by lrint

lrint is at least as fast, avoids an implicit cast, and uses a superior
rounding mode.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavc/libvpxenc: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:48:45 +0000 (14:48 -0500)]
lavc/libvpxenc: replace round by lrint

Mostly cosmetic here.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/vf_drawtext: replace round by llrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:39:37 +0000 (14:39 -0500)]
lavfi/vf_drawtext: replace round by llrint

llrint is at least as fast, and avoids an implicit cast.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/vf_colorlevels: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:36:53 +0000 (14:36 -0500)]
lavfi/vf_colorlevels: replace round by lrint

lrint avoids an implicit cast, and is not slower on non-broken libm's. Thus this
represents a Pareto improvement.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/vf_colorchannelmixer: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:34:09 +0000 (14:34 -0500)]
lavfi/vf_colorchannelmixer: replace round by lrint

lrint is faster here on -ftree-vectorize with GCC. This is likely simply
an artifact of GCC's rather terrible auto-vectorizer, since as per the
instruction set manuals cvtsd2si and cvttsd2si (or their vector equivalents)
have identical cycle timings.

Anyway, regardless of above, lrint is superior to round accuracy wise.

Safety guaranteed as long int has at least 32 bits.

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/drawtext: fix shadow[xy] descriptions
Clément Bœsch [Mon, 21 Dec 2015 15:42:14 +0000 (16:42 +0100)]
lavfi/drawtext: fix shadow[xy] descriptions

4 years agolavfi/drawtext: hide first font load warning when fontconfig is present
Clément Bœsch [Mon, 21 Dec 2015 15:07:22 +0000 (16:07 +0100)]
lavfi/drawtext: hide first font load warning when fontconfig is present

4 years agolavfi/drawtext: fix crash when no text, file or timecode provided
Clément Bœsch [Mon, 21 Dec 2015 14:54:20 +0000 (15:54 +0100)]
lavfi/drawtext: fix crash when no text, file or timecode provided

4 years agoavfilter/af_ladspa: fix av_assert0()
Paul B Mahol [Mon, 21 Dec 2015 12:26:17 +0000 (13:26 +0100)]
avfilter/af_ladspa: fix av_assert0()

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agolavf/img2enc: add atomic_writing option
Clément Bœsch [Tue, 8 Dec 2015 09:44:31 +0000 (10:44 +0100)]
lavf/img2enc: add atomic_writing option

This behaviour change caused a regression on our side recently, we might
want to disable the option by default.

4 years agolavc: add text encoder
Clément Bœsch [Thu, 17 Dec 2015 11:23:35 +0000 (12:23 +0100)]
lavc: add text encoder

4 years agolavfi/scale: add nb_slices debug option
Clément Bœsch [Fri, 18 Dec 2015 14:44:33 +0000 (15:44 +0100)]
lavfi/scale: add nb_slices debug option

4 years agoavcodec/indeo2: use init_get_bits8
Paul B Mahol [Sun, 20 Dec 2015 20:31:55 +0000 (21:31 +0100)]
avcodec/indeo2: use init_get_bits8

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agox86/hevc_sao: add ff_hevc_sao_edge_filter_{8,16}_{10,12}
James Almer [Sun, 6 Dec 2015 05:47:45 +0000 (02:47 -0300)]
x86/hevc_sao: add ff_hevc_sao_edge_filter_{8,16}_{10,12}

Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
4 years agox86/hevc_sao: simplify sao_edge_filter 10/12bit
James Almer [Sun, 6 Dec 2015 05:46:51 +0000 (02:46 -0300)]
x86/hevc_sao: simplify sao_edge_filter 10/12bit

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
4 years agox86/hevc_sao: simplify sao_band_filter 10/12bit
James Almer [Sun, 6 Dec 2015 05:42:34 +0000 (02:42 -0300)]
x86/hevc_sao: simplify sao_band_filter 10/12bit

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
4 years agoavfilter/avf_showfreqs: make it possible to split channels
Paul B Mahol [Sun, 20 Dec 2015 18:52:51 +0000 (19:52 +0100)]
avfilter/avf_showfreqs: make it possible to split channels

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavcodec/flacenc: use designated initializers for AVClass
Paul B Mahol [Sun, 20 Dec 2015 16:47:21 +0000 (17:47 +0100)]
avcodec/flacenc: use designated initializers for AVClass

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavcodec/s302menc: check if buf_size can actually be put into 16bit size
Paul B Mahol [Sat, 19 Dec 2015 20:52:19 +0000 (21:52 +0100)]
avcodec/s302menc: check if buf_size can actually be put into 16bit size

This disallows creating unplayable audio.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavcodec/s302menc: set supported channel layouts by codec
Paul B Mahol [Sat, 19 Dec 2015 20:34:27 +0000 (21:34 +0100)]
avcodec/s302menc: set supported channel layouts by codec

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agonuv: sanitize negative fps rate
Andreas Cadhalpun [Wed, 16 Dec 2015 19:52:39 +0000 (20:52 +0100)]
nuv: sanitize negative fps rate

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agorawdec: only exempt BIT0 with need_copy from buffer sanity check
Andreas Cadhalpun [Sat, 19 Dec 2015 22:45:06 +0000 (23:45 +0100)]
rawdec: only exempt BIT0 with need_copy from buffer sanity check

Otherwise the too small buffer is directly used in the frame, causing
segmentation faults, when trying to use the frame.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agomlvdec: check that index_entries exist
Andreas Cadhalpun [Sat, 19 Dec 2015 22:44:53 +0000 (23:44 +0100)]
mlvdec: check that index_entries exist

This fixes NULL pointer dereferencing.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agoavcodec/mpeg4videodec: also for empty partitioned slices
Michael Niedermayer [Sat, 19 Dec 2015 22:21:33 +0000 (23:21 +0100)]
avcodec/mpeg4videodec: also for empty partitioned slices

Fixes assertion failure
Fixes: id_acf3e47f864e1ee4c7b86c0653e0ff31e5bde56e.m4v

Found-by: Andreas Cadhalpun <andreas.cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agolavu/libm: add copysign hack
Ganesh Ajjanagadde [Fri, 18 Dec 2015 19:13:11 +0000 (11:13 -0800)]
lavu/libm: add copysign hack

For systems with broken libms.
Tested with NAN, -NAN, INFINITY, -INFINITY, +/-x for regular double x and
combinations of these.

Old versions of MSVC need some UINT64_C hackery.

Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Reviewed-by: Hendrik Leppkes <h.leppkes@gmail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agoavcodec/Makefile: add missing dep for g723_1 encoder
James Almer [Sat, 19 Dec 2015 21:50:47 +0000 (18:50 -0300)]
avcodec/Makefile: add missing dep for g723_1 encoder

Signed-off-by: James Almer <jamrial@gmail.com>
4 years agoavfilter/af_dynaudnorm: use av_malloc_array()
Paul B Mahol [Sat, 19 Dec 2015 21:46:10 +0000 (22:46 +0100)]
avfilter/af_dynaudnorm: use av_malloc_array()

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavcodec/h264_refs: Fix long_idx check
Michael Niedermayer [Sat, 19 Dec 2015 20:59:42 +0000 (21:59 +0100)]
avcodec/h264_refs: Fix long_idx check

Fixes out of array read
Fixes mozilla bug 1233606

Found-by: Tyson Smith
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoswscale/arm/yuv2rgb: add ff_yuv420p_to_{argb,rgba,abgr,bgra}_neon_{16,32}
Matthieu Bouron [Tue, 15 Dec 2015 13:42:22 +0000 (14:42 +0100)]
swscale/arm/yuv2rgb: add ff_yuv420p_to_{argb,rgba,abgr,bgra}_neon_{16,32}

4 years agoswscale/arm/yuv2rgb: disable neon if accurate_rnd is enabled
Matthieu Bouron [Fri, 18 Dec 2015 13:24:52 +0000 (14:24 +0100)]
swscale/arm/yuv2rgb: disable neon if accurate_rnd is enabled

This disables the 32bit precision neon code path in favor of the
default C one and avoids breaking fate.

4 years agoavcodec/ffv1enc: Fix 2 pass mode with the default rc table
Michael Niedermayer [Sat, 19 Dec 2015 19:06:58 +0000 (20:06 +0100)]
avcodec/ffv1enc: Fix 2 pass mode with the default rc table

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoavfilter/vf_stereo3d: add interleave columns input support
Paul B Mahol [Fri, 18 Dec 2015 21:00:31 +0000 (22:00 +0100)]
avfilter/vf_stereo3d: add interleave columns input support

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agolavu/intmath: add faster clz support
Ganesh Ajjanagadde [Wed, 16 Dec 2015 18:28:39 +0000 (13:28 -0500)]
lavu/intmath: add faster clz support

This should be useful for the sofalizer filter.

Reviewed-by: Kieran Kunhya <kierank@ob-encoder.com>
Reviewed-by: Clément Bœsch <u@pkh.me>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/vsrc_mandelbrot: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:54:53 +0000 (14:54 -0500)]
lavfi/vsrc_mandelbrot: replace round by lrint

lrint is at least as fast, and is more accurate.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/vf_cropdetect: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:52:16 +0000 (14:52 -0500)]
lavfi/vf_cropdetect: replace round by lrint

lrint is at least as fast, and more accurate.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavf/hlsenc: replace round by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 19:40:50 +0000 (14:40 -0500)]
lavf/hlsenc: replace round by lrint

Mainly cosmetic here.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavfi/vf_idet: replace round and cast by lrint
Ganesh Ajjanagadde [Wed, 16 Dec 2015 18:03:59 +0000 (13:03 -0500)]
lavfi/vf_idet: replace round and cast by lrint

lrint is faster and conveys the intent better here. It is safe as long int has
at least 32 bits.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agolavc/aacsbr: sbr_dequant optimization
Ganesh Ajjanagadde [Wed, 16 Dec 2015 04:27:23 +0000 (23:27 -0500)]
lavc/aacsbr: sbr_dequant optimization

This uses ff_exp2fi to get a speedup (~ 6x).

sample benchmark (Haswell, GNU/Linux):
old:
  19102 decicycles in sbr_dequant,    1023 runs,      1 skips
  19002 decicycles in sbr_dequant,    2045 runs,      3 skips
  17638 decicycles in sbr_dequant,    4093 runs,      3 skips
  15825 decicycles in sbr_dequant,    8189 runs,      3 skips
  16404 decicycles in sbr_dequant,   16379 runs,      5 skips

new:
   3063 decicycles in sbr_dequant,    1024 runs,      0 skips
   3049 decicycles in sbr_dequant,    2048 runs,      0 skips
   2968 decicycles in sbr_dequant,    4096 runs,      0 skips
   2818 decicycles in sbr_dequant,    8191 runs,      1 skips
   2853 decicycles in sbr_dequant,   16383 runs,      1 skips

Reviewed-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
4 years agonutdec: reject negative value_len in read_sm_data
Andreas Cadhalpun [Sat, 19 Dec 2015 11:02:56 +0000 (12:02 +0100)]
nutdec: reject negative value_len in read_sm_data

If it is negative, it can cause the byte position to move backwards in
avio_skip, which in turn makes sm_size negative and thus size larger
than the size of the packet buffer, causing invalid writes in avio_read.

Also fix potential overflow of avio_tell(bc) + value_len.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agoxwddec: prevent overflow of lsize * avctx->height
Andreas Cadhalpun [Fri, 18 Dec 2015 18:28:51 +0000 (19:28 +0100)]
xwddec: prevent overflow of lsize * avctx->height

This is used to check if the input buffer is large enough, so if this
overflows it can cause a false negative leading to a segmentation fault
in bytestream2_get_bufferu.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agoffplay: remove existing AVPicture usage
Marton Balint [Tue, 8 Dec 2015 21:33:31 +0000 (22:33 +0100)]
ffplay: remove existing AVPicture usage

It is deprecated.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>
4 years agonutdec: only copy the header if it exists
Andreas Cadhalpun [Fri, 18 Dec 2015 14:18:47 +0000 (15:18 +0100)]
nutdec: only copy the header if it exists

Fixes ubsan runtime error: null pointer passed as argument 2, which is
declared to never be null

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agoffserver: refactor build_feed_streams()
Reynaldo H. Verdejo Pinochet [Thu, 17 Dec 2015 02:14:25 +0000 (18:14 -0800)]
ffserver: refactor build_feed_streams()

* Avoid excesive nesting that made it really hard to follow
* Drop unneeded vars
* Factor out codec compatibility check routine
* Ensure inputs are closed and contexts are freed as needed
  before returning

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
4 years agoffserver: refactor build_file_streams()
Reynaldo H. Verdejo Pinochet [Wed, 16 Dec 2015 06:08:52 +0000 (22:08 -0800)]
ffserver: refactor build_file_streams()

Avoid unneeded nesting, drop redundant var

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
4 years agoffserver: unify exit path from build_feed_streams()
Reynaldo H. Verdejo Pinochet [Tue, 15 Dec 2015 22:51:18 +0000 (14:51 -0800)]
ffserver: unify exit path from build_feed_streams()

Exit from main on build_feed_streams() failures & use
standard EXIT_ codes on error out/normal exit.

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>
4 years agoswscale/arm/yuv2rgb: simplify process_16px_* macro call
Matthieu Bouron [Tue, 15 Dec 2015 16:04:09 +0000 (17:04 +0100)]
swscale/arm/yuv2rgb: simplify process_16px_* macro call

4 years agoavcodec/arm64: fix inverted register order in transpose_4x4H
Janne Grunau [Fri, 18 Dec 2015 10:27:05 +0000 (11:27 +0100)]
avcodec/arm64: fix inverted register order in transpose_4x4H

Fix related register order issue in ff_h264_idct_add_neon.

Found-by: zjh8890 <243186085@qq.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoavfilter/vf_delogo: change the definition of logo_x2 and logo_y2
Jean Delvare [Fri, 18 Dec 2015 15:16:38 +0000 (16:16 +0100)]
avfilter/vf_delogo: change the definition of logo_x2 and logo_y2

In the code we keep using logo_x2-1 and logo_y2-1 rather than logo_x2
and logo_y2 themselves. Define them to be what we need instead, to avoid
the repeated subtractions.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agooggparsedaala: sync with current bitstream syntax
Rostislav Pehlivanov [Fri, 18 Dec 2015 22:54:38 +0000 (22:54 +0000)]
oggparsedaala: sync with current bitstream syntax

Since the parser was merged back almost 2 months ago this is the first
time the bitstream of the container has been updated.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
4 years agoavfilter/vf_stereo3d: fix interleaved columns output
Paul B Mahol [Fri, 18 Dec 2015 21:04:19 +0000 (22:04 +0100)]
avfilter/vf_stereo3d: fix interleaved columns output

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavfilter/vf_stereo3d: multiply linesize only once for interleaved row to mono
Paul B Mahol [Fri, 18 Dec 2015 18:44:01 +0000 (19:44 +0100)]
avfilter/vf_stereo3d: multiply linesize only once for interleaved row to mono

Signed-off-by: Paul B Mahol <onemda@gmail.com>
4 years agoavcodec/ffv1enc: unbreak -coder option
Michael Niedermayer [Fri, 18 Dec 2015 16:52:35 +0000 (17:52 +0100)]
avcodec/ffv1enc: unbreak -coder option

This fixes a segfault caused by moving the coder option and changing its semantics

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
4 years agoswscale/arm/yuv2rgb: fix slicing
Clément Bœsch [Fri, 18 Dec 2015 15:03:00 +0000 (16:03 +0100)]
swscale/arm/yuv2rgb: fix slicing

4 years agoexr: fix out of bounds read in get_code
Andreas Cadhalpun [Sun, 13 Dec 2015 22:17:09 +0000 (23:17 +0100)]
exr: fix out of bounds read in get_code

This macro unconditionally used out[-1], which causes an out of bounds
read, if out is the very beginning of the buffer.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
4 years agoon2avc: limit number of bits to 30 in get_egolomb
Andreas Cadhalpun [Wed, 16 Dec 2015 15:48:19 +0000 (16:48 +0100)]
on2avc: limit number of bits to 30 in get_egolomb

More don't fit into the integer output.

Also use get_bits_long, since get_bits only supports reading up to 25
bits, while get_bits_long supports the full integer range.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>