ffmpeg.git
23 months agoavformat/movenc: Add support for more colorspaces
Steven Robertson [Thu, 14 Dec 2017 19:52:45 +0000 (11:52 -0800)]
avformat/movenc: Add support for more colorspaces

With FCPX 10.4, Apple has expanded the set of colorspace, primaries,
and trc flags officially supported in QuickTime files. The expanded set
matches the codepoints used in ffmpeg and many other specs.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavcodec/vp9: mark frame as finished on decode_tiles() failure
Michael Niedermayer [Thu, 14 Dec 2017 01:02:55 +0000 (02:02 +0100)]
avcodec/vp9: mark frame as finished on decode_tiles() failure

Fixes deadlock with framethreads
Fixes: Netflix_Aerial_1080p_60fps_8bit_420.y4m.vp9.noaltref.webm.ivf.s69372_r01-05_b6-.ivf
Fixes: Netflix_Aerial_1080p_60fps_10bit_420.y4m.vp9.noaltref.webm.ivf.s149104_r01-05_b6-.ivf
Fixes: ducks_take_off_444_720p50.y4m.vp9.webm.ivf.s107375_r01-05_b6-.ivf

Reported-by: James Zern <jzern@google.com>
Reviewed-by: James Zern <jzern@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavfilter/drawbox+drawgrid - add option to prevent overwriting of source pixels
Gyan Doshi [Mon, 11 Dec 2017 17:05:18 +0000 (22:35 +0530)]
avfilter/drawbox+drawgrid - add option to prevent overwriting of source pixels

If the user-supplied color in drawbox and drawgrid filters is non-opaque,
the box & grid painting overwrites the input's pixels (including alpha).
Users typically expect the alpha of the specified color to only act as a key
for compositing on top of the main input.

Added option allows users to select between replacement and composition.
Tested and documented.

23 months agoMerge commit '47687a2f8aca3f65b6fdd117b1cb66a7409a7fd1'
James Almer [Thu, 14 Dec 2017 22:10:12 +0000 (19:10 -0300)]
Merge commit '47687a2f8aca3f65b6fdd117b1cb66a7409a7fd1'

* commit '47687a2f8aca3f65b6fdd117b1cb66a7409a7fd1':
  avcodec: add metadata to identify wrappers and hardware decoders

This commit is a noop, see b945fed629a872d393f59d16fc5773574126ca88

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit '508378556631dc18d32247b4a4e35703758e1ca9'
James Almer [Thu, 14 Dec 2017 22:04:37 +0000 (19:04 -0300)]
Merge commit '508378556631dc18d32247b4a4e35703758e1ca9'

* commit '508378556631dc18d32247b4a4e35703758e1ca9':
  qsv: Support explicit lookahead downscaling

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit 'f2e9a0ecbef5027f9532c49ffcdfc11d199f6150'
James Almer [Thu, 14 Dec 2017 21:46:43 +0000 (18:46 -0300)]
Merge commit 'f2e9a0ecbef5027f9532c49ffcdfc11d199f6150'

* commit 'f2e9a0ecbef5027f9532c49ffcdfc11d199f6150':
  qsv/vp8dec: fixes memory leak issue

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit '7c65a76b16bc3a44f1592acde2176f187a058797'
James Almer [Thu, 14 Dec 2017 21:46:10 +0000 (18:46 -0300)]
Merge commit '7c65a76b16bc3a44f1592acde2176f187a058797'

* commit '7c65a76b16bc3a44f1592acde2176f187a058797':
  lavc/qsvenc: add error messeage if ICQ unsupported.

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit 'bddb2ce179c57db6e3c79fdc3363c165d90850b0'
James Almer [Thu, 14 Dec 2017 21:45:20 +0000 (18:45 -0300)]
Merge commit 'bddb2ce179c57db6e3c79fdc3363c165d90850b0'

* commit 'bddb2ce179c57db6e3c79fdc3363c165d90850b0':
  lavc/qsvenc: ICQ/VCM/QVBR are not avilable on Linux

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit '6ff29343b01923e9b125fe7404ac8701cdfb1fe5'
James Almer [Thu, 14 Dec 2017 21:44:42 +0000 (18:44 -0300)]
Merge commit '6ff29343b01923e9b125fe7404ac8701cdfb1fe5'

* commit '6ff29343b01923e9b125fe7404ac8701cdfb1fe5':
  lavc/qsvenc: set HRD buffer size

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit '18a0f420269ff4c730422361c5c4d8eea096e900'
James Almer [Thu, 14 Dec 2017 21:44:04 +0000 (18:44 -0300)]
Merge commit '18a0f420269ff4c730422361c5c4d8eea096e900'

* commit '18a0f420269ff4c730422361c5c4d8eea096e900':
  checkasm: Use LOCAL_ALIGNED for aligned variables on the stack

This commit is a noop, see
https://ffmpeg.org/pipermail/ffmpeg-devel/2015-December/184777.html

Merged-by: James Almer <jamrial@gmail.com>
23 months agoavcodec: add metadata to identify wrappers and hardware decoders
wm4 [Mon, 11 Dec 2017 15:18:44 +0000 (16:18 +0100)]
avcodec: add metadata to identify wrappers and hardware decoders

Explicitly identify decoder/encoder wrappers with a common name. This
saves API users from guessing by the name suffix. For example, they
don't have to guess that "h264_qsv" is the h264 QSV implementation, and
instead they can just check the AVCodec .codec and .wrapper_name fields.

Explicitly mark AVCodec entries that are hardware decoders or most
likely hardware decoders with new AV_CODEC_CAPs. The purpose is allowing
API users listing hardware decoders in a more generic way. The proposed
AVCodecHWConfig does not provide this information fully, because it's
concerned with decoder configuration, not information about the fact
whether the hardware is used or not.

AV_CODEC_CAP_HYBRID exists specifically for QSV, which can have software
implementations in case the hardware is not capable.

Based on a patch by Philip Langdale <philipl@overt.org>.

Merges Libav commit 47687a2f8aca3f65b6fdd117b1cb66a7409a7fd1.

23 months agoavcodec: add metadata to identify wrappers and hardware decoders
wm4 [Tue, 12 Dec 2017 10:03:43 +0000 (11:03 +0100)]
avcodec: add metadata to identify wrappers and hardware decoders

Explicitly identify decoder/encoder wrappers with a common name. This
saves API users from guessing by the name suffix. For example, they
don't have to guess that "h264_qsv" is the h264 QSV implementation, and
instead they can just check the AVCodec .codec and .wrapper_name fields.

Explicitly mark AVCodec entries that are hardware decoders or most
likely hardware decoders with new AV_CODEC_CAPs. The purpose is allowing
API users listing hardware decoders in a more generic way. The proposed
AVCodecHWConfig does not provide this information fully, because it's
concerned with decoder configuration, not information about the fact
whether the hardware is used or not.

AV_CODEC_CAP_HYBRID exists specifically for QSV, which can have software
implementations in case the hardware is not capable.

Based on a patch by Philip Langdale <philipl@overt.org>.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
23 months agolavf/mpegts: mark packets with TEI flag as corrupted
Rodger Combs [Thu, 19 Oct 2017 06:58:10 +0000 (01:58 -0500)]
lavf/mpegts: mark packets with TEI flag as corrupted

23 months agolibavcodec/decode: remove duplicate includes
Michael Niedermayer [Wed, 13 Dec 2017 20:51:31 +0000 (21:51 +0100)]
libavcodec/decode: remove duplicate includes

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoivfenc: add AV1 support
Tristan Matthews [Sun, 10 Dec 2017 23:28:50 +0000 (18:28 -0500)]
ivfenc: add AV1 support

libaom tools work with ivf files.

23 months agopatcheck: Add 'threshhold' to common typo list
Kelly Ledford [Tue, 12 Dec 2017 19:31:24 +0000 (11:31 -0800)]
patcheck: Add 'threshhold' to common typo list

Signed-off-by: Kelly Ledford <kelly.ledford@intel.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agolibavfilter/af_dcshift.c: Fixed repeated spelling error
Kelly Ledford [Tue, 12 Dec 2017 19:31:23 +0000 (11:31 -0800)]
libavfilter/af_dcshift.c: Fixed repeated spelling error

'threshhold' should be 'threshold'

Signed-off-by: Kelly Ledford <kelly.ledford@intel.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agocheckasm/vf_hflip : add test for vf_hflip byte and short simd
Martin Vignali [Mon, 11 Dec 2017 10:26:44 +0000 (11:26 +0100)]
checkasm/vf_hflip : add test for vf_hflip byte and short simd

23 months agoavfilter/vf_hflip : move context func init in ff_hflip_init
Martin Vignali [Mon, 11 Dec 2017 10:22:56 +0000 (11:22 +0100)]
avfilter/vf_hflip : move context func init in ff_hflip_init

23 months agolibavcodec/mpegvideo_parser: improve detection of progressive mpeg2
Aman Gupta [Wed, 6 Dec 2017 22:54:12 +0000 (14:54 -0800)]
libavcodec/mpegvideo_parser: improve detection of progressive mpeg2

Previously many progressive mpeg2video samples were detected as interlaced
by ffmpeg/ffprobe. For example, https://tmm1.s3.amazonaws.com/720p.ts

Before:

    Input #0, mpegts, from 'https://tmm1.s3.amazonaws.com/720p.ts':
      Duration: 00:00:08.62, start: 18974.073233, bitrate: 9734 kb/s
      Program 2
     Stream #0:0[0x12eb]: Video: mpeg2video (Main), yuv420p(tv, bottom first), 1280x720 ...

After:

    Input #0, mpegts, from 'https://tmm1.s3.amazonaws.com/720p.ts':
      Duration: 00:00:08.62, start: 18974.073233, bitrate: 9734 kb/s
      Program 2
     Stream #0:0[0x12eb]: Video: mpeg2video (Main), yuv420p(tv, progressive), 1280x720 ...

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavfilter/formats: fix wrong function name in error message
Jun Zhao [Mon, 4 Dec 2017 04:50:34 +0000 (12:50 +0800)]
avfilter/formats: fix wrong function name in error message

Use perdefined micro __FUNCTION__ rather than hard coding function name
to fix wrong function name in error message.

Signed-off-by: Jun Zhao <jun.zhao@intel.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavformat/mux: factorize AVFormatContext->avoid_negative_ts initialization
James Almer [Mon, 11 Dec 2017 04:02:30 +0000 (01:02 -0300)]
avformat/mux: factorize AVFormatContext->avoid_negative_ts initialization

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>
23 months agoavfilter/vf_pseudocolor: add support for more formats
Paul B Mahol [Tue, 12 Dec 2017 18:31:50 +0000 (19:31 +0100)]
avfilter/vf_pseudocolor: add support for more formats

Signed-off-by: Paul B Mahol <onemda@gmail.com>
23 months agoqsv: Support explicit lookahead downscaling
Luca Barbato [Tue, 12 Dec 2017 14:27:52 +0000 (14:27 +0000)]
qsv: Support explicit lookahead downscaling

23 months agoqsv/vp8dec: fixes memory leak issue
Li, Zhong [Thu, 7 Dec 2017 07:26:38 +0000 (15:26 +0800)]
qsv/vp8dec: fixes memory leak issue

Signed-off-by: Zhong Li <zhong.li@intel.com>
Signed-off-by: Maxym Dmytrychenko <maxim.d33@gmail.com>
23 months agolavc/qsvenc: add error messeage if ICQ unsupported.
Li, Zhong [Thu, 7 Dec 2017 06:44:27 +0000 (14:44 +0800)]
lavc/qsvenc: add error messeage if ICQ unsupported.

Signed-off-by: Zhong Li <zhong.li@intel.com>
Signed-off-by: Maxym Dmytrychenko <maxim.d33@gmail.com>
23 months agolavc/qsvenc: ICQ/VCM/QVBR are not avilable on Linux
Li, Zhong [Thu, 7 Dec 2017 06:44:26 +0000 (14:44 +0800)]
lavc/qsvenc: ICQ/VCM/QVBR are not avilable on Linux

Signed-off-by: Zhong Li <zhong.li@intel.com>
Signed-off-by: Maxym Dmytrychenko <maxim.d33@gmail.com>
23 months agolavc/qsvenc: set HRD buffer size
Li, Zhong [Fri, 1 Dec 2017 08:48:45 +0000 (16:48 +0800)]
lavc/qsvenc: set HRD buffer size

Hypothetical Reference Decoding (HRD) model assumes that data flows
into a buffer of the fixed size BufferSizeInKB with a constant bitrate.
Smaller BufferSizeInKB means smaller frame size variations,
but more difficult to maintain HRD.

Signed-off-by: Zhong Li <zhong.li@intel.com>
Signed-off-by: Maxym Dmytrychenko <maxim.d33@gmail.com>
23 months agocheckasm: Use LOCAL_ALIGNED for aligned variables on the stack
Martin Storsjö [Mon, 11 Dec 2017 20:25:16 +0000 (22:25 +0200)]
checkasm: Use LOCAL_ALIGNED for aligned variables on the stack

This fixes fate-checkasm-hevc_mc on ARMCC 5.0 after adding
NEON HEVC MC assembly.

Signed-off-by: Martin Storsjö <martin@martin.st>
23 months agoavformat/hlsenc: reindent after previous commits
Steven Liu [Tue, 12 Dec 2017 03:50:23 +0000 (11:50 +0800)]
avformat/hlsenc: reindent after previous commits

23 months agodoc/filters: re-arrange options for testsrc family
Lou Logan [Mon, 11 Dec 2017 21:46:02 +0000 (12:46 -0900)]
doc/filters: re-arrange options for testsrc family

Additionally:
* Mention that allrgb and allyuv do not support the "size" option.
* Separate examples into subsection.

Fixes ticket #6906.

Signed-off-by: Lou Logan <lou@lrcd.com>
23 months agoavfilter/avfiltergraph: remove ugly dead code
Paul B Mahol [Mon, 11 Dec 2017 21:39:26 +0000 (22:39 +0100)]
avfilter/avfiltergraph: remove ugly dead code

Remnant of old merge.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
23 months agoRevert "libavcodec/utils.c: simplify avcodec locking with atomics"
Hendrik Leppkes [Mon, 11 Dec 2017 19:59:56 +0000 (20:59 +0100)]
Revert "libavcodec/utils.c: simplify avcodec locking with atomics"

This reverts commit 590136e78da3d091ea99ab5432543d47a559a461.

Atomics are not required for this variable, because it is protected
through the lock manager, and the use of atomics here is not compatible
with the c11 emulation wrappersi.

Fixes FATE on MSVC, among other setups which use the compat wrappers.

23 months agoavcodec/libvpx: remove disabled code
James Almer [Mon, 11 Dec 2017 03:42:26 +0000 (00:42 -0300)]
avcodec/libvpx: remove disabled code

Signed-off-by: James Almer <jamrial@gmail.com>
23 months agoavformat/hlsenc: move init operations from write_header to init
Steven Liu [Mon, 11 Dec 2017 03:03:25 +0000 (11:03 +0800)]
avformat/hlsenc: move init operations from write_header to init

Reviewed-by: Vishwanath Dixit <vdixit@akamai.com>
Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
23 months agoh264dec: Remove mpeg4video.h header dependency
Kieran Kunhya [Sun, 10 Dec 2017 23:51:37 +0000 (23:51 +0000)]
h264dec: Remove mpeg4video.h header dependency

23 months agolibavcodec: Move ff_print_debug_info2 to mpegutils.c
Kieran Kunhya [Mon, 11 Dec 2017 00:04:44 +0000 (00:04 +0000)]
libavcodec: Move ff_print_debug_info2 to mpegutils.c

23 months agodoc/libav-merge: add a line about the skipped HEVC MC arm functions
James Almer [Mon, 11 Dec 2017 00:03:28 +0000 (21:03 -0300)]
doc/libav-merge: add a line about the skipped HEVC MC arm functions

Signed-off-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit '7993ec19af394fdc58ec64165bc0b12619543a5d'
James Almer [Mon, 11 Dec 2017 00:02:45 +0000 (21:02 -0300)]
Merge commit '7993ec19af394fdc58ec64165bc0b12619543a5d'

* commit '7993ec19af394fdc58ec64165bc0b12619543a5d':
  hevc: Add hevc_get_pixel_4/8/12/16/24/32/48/64

This commit is a noop, see
https://ffmpeg.org/pipermail/ffmpeg-devel/2015-December/184777.html

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit '3ff8fbbf5a7bc40c09db74d4952364997fd3c611'
James Almer [Sun, 10 Dec 2017 23:54:49 +0000 (20:54 -0300)]
Merge commit '3ff8fbbf5a7bc40c09db74d4952364997fd3c611'

* commit '3ff8fbbf5a7bc40c09db74d4952364997fd3c611':
  vaapi_h265: Add named options for setting profile and level
  vaapi_h264: Add named options for setting profile and level
  vaapi: Make the decode profile matching more explicit
  vaapi_h264: Fix VUI max_dec_frame_buffering
  vaapi_h265: Enable VBR mode

This commit is a noop, see
385cafb07ac1e46433931ea9749a134efd7350be
f31478ba1472afe5c1eed60f219ae331816425a2
efd0612fdcb7490ed371899f532d73ef8bb7cba0
c490fc9536dcea7fdf1245a340bf075533610bc2
71421f382f5c53b1a109489a906a80c1b7e3eed7

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit '5a6707e49b7710f48d658b2f2591b9a6337fb9b7'
James Almer [Sun, 10 Dec 2017 23:49:50 +0000 (20:49 -0300)]
Merge commit '5a6707e49b7710f48d658b2f2591b9a6337fb9b7'

* commit '5a6707e49b7710f48d658b2f2591b9a6337fb9b7':
  cbs_mpeg2: Fix marker_bit type
  cbs: Add padding to slice data allocations

This commit is a noop, see
b4c915f4b3e15c3e787e319b961e4389762f6309
5b2c71bb94d7cab23ee81b5c29388f5fadbcaf22

Merged-by: James Almer <jamrial@gmail.com>
23 months agoMerge commit 'f6161fccf8c5720ceac1ed1df8ba60ff8fed69f5'
James Almer [Sun, 10 Dec 2017 23:44:26 +0000 (20:44 -0300)]
Merge commit 'f6161fccf8c5720ceac1ed1df8ba60ff8fed69f5'

* commit 'f6161fccf8c5720ceac1ed1df8ba60ff8fed69f5':
  rtsp: only break on parse_rtsp_message on error

Merged-by: James Almer <jamrial@gmail.com>
23 months agolavc/huffyuvenc: Move a variable declaration up.
Carl Eugen Hoyos [Sun, 10 Dec 2017 19:46:55 +0000 (20:46 +0100)]
lavc/huffyuvenc: Move a variable declaration up.

Fixes a warning with gcc-6.3:
libavcodec/huffyuvenc.c:93:5: warning: ISO C90 forbids mixed declarations and code

23 months agoavcodec/vp9_superframe_split_bsf: Fix integer overflow in frame_size/total_size checks
Michael Niedermayer [Wed, 6 Dec 2017 21:42:05 +0000 (22:42 +0100)]
avcodec/vp9_superframe_split_bsf: Fix integer overflow in frame_size/total_size checks

Fixes: signed integer overflow: -1698586465 + -551542752 cannot be represented in type 'int'
Fixes: 4490/clusterfuzz-testcase-minimized-5210014592532480

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavcodec/amrwbdec: Fix division by 0 in voice_factor()
Michael Niedermayer [Thu, 7 Dec 2017 14:32:54 +0000 (15:32 +0100)]
avcodec/amrwbdec: Fix division by 0 in voice_factor()

The added value matches "Digital cellular telecommunications system (Phase 2+) (GSM); Universal Mobile Telecommunications System (UMTS); LTE; Extended Adaptive Multi-Rate - Wideband (AMR-WB+) codec; Floating-point ANSI-C code (3GPP TS 26.304 version 14.0.0 Release 14)
Extended Adaptive Multi-Rate - Wideband (AMR-WB+) codec; Floating-point ANSI-C code"

Fixes: runtime error: division by zero
Fixes: 4415/clusterfuzz-testcase-minimized-4677752314658816

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoconfigure: Move V4L2 M2M help line to the hardware library section
Mark Thompson [Sun, 10 Dec 2017 17:15:06 +0000 (17:15 +0000)]
configure: Move V4L2 M2M help line to the hardware library section

23 months agolavc: Make hardware config method support more explicit for hwaccels
Thomas Guillem [Fri, 8 Dec 2017 15:59:39 +0000 (15:59 +0000)]
lavc: Make hardware config method support more explicit for hwaccels

This fixes the use of old ad-hoc methods which are still supported by some
hwaccels which also support newer methods (DXVA2, VAAPI, VDPAU,
videotoolbox) - without the method being visible here, ff_get_format()
would refuse to use it.

Signed-off-by: Mark Thompson <sw@jkqxz.net>
23 months agofix MSVC compilation errors
Mateusz [Fri, 8 Dec 2017 07:46:52 +0000 (08:46 +0100)]
fix MSVC compilation errors

After commit 3701d49 'error_resilience: remove avpriv_atomic usage'
we have included windows.h in much more files and we should
avoid conflicts with defines/function declarations.

Signed-off-by: Mateusz Brzostek <mateuszb@poczta.onet.pl>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agocheckasm/llviddsp: fix mixed code and declarations
James Almer [Sun, 10 Dec 2017 03:51:35 +0000 (00:51 -0300)]
checkasm/llviddsp: fix mixed code and declarations

Signed-off-by: James Almer <jamrial@gmail.com>
23 months agox86/lossless_videodsp: rename ff_add_left_pred_int16_sse4 to ff_add_left_pred_int16_u...
James Almer [Sun, 10 Dec 2017 03:51:01 +0000 (00:51 -0300)]
x86/lossless_videodsp: rename ff_add_left_pred_int16_sse4 to ff_add_left_pred_int16_unaligned_ssse3

SSSE3_FAST is the proper check for it.

Signed-off-by: James Almer <jamrial@gmail.com>
23 months agox86/lossless_videodsp: don't overread the dst buffer in ff_add_left_pred_unaligned_avx2
James Almer [Sun, 10 Dec 2017 03:38:05 +0000 (00:38 -0300)]
x86/lossless_videodsp: don't overread the dst buffer in ff_add_left_pred_unaligned_avx2

Fixes valgrind

Signed-off-by: James Almer <jamrial@gmail.com>
23 months agoamf: fix wrong profile level after auto-correction in H264 and HEVC
Mironov, Mikhail [Fri, 8 Dec 2017 19:23:38 +0000 (19:23 +0000)]
amf: fix wrong profile level after auto-correction in H264 and HEVC

Moved bitrate parameters set before Init() call because bitrate is used in
profile level correction code inside Init().

Signed-off-by: Mikhail Mironov <mikhail.mironov@amd.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agolavf/mov: modify code indentation
tiejun.peng [Fri, 8 Dec 2017 16:39:29 +0000 (00:39 +0800)]
lavf/mov: modify code indentation

Signed-off-by: tiejun.peng <tiejun.peng@foxmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agolavf/mov: add some useful warning log of eof
tiejun.peng [Fri, 8 Dec 2017 16:28:41 +0000 (00:28 +0800)]
lavf/mov: add some useful warning log of eof

Signed-off-by: tiejun.peng <tiejun.peng@foxmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavcodec/Makefile: Fix opus parser dependency.
Jacob Trimble [Thu, 7 Dec 2017 19:05:46 +0000 (11:05 -0800)]
avcodec/Makefile: Fix opus parser dependency.

The opus.c file uses ff_celt_freq_range and ff_celt_freq_bands which are
defined in opustab.c.  The opus parser needs to include that file to
avoid linker errors when not including the opus encoder/decoder.

Signed-off-by: Jacob Trimble <modmaker@google.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavcodec/decode: reset codec on receiving packet after EOF in compat_decode
James Cowgill [Thu, 9 Nov 2017 12:21:23 +0000 (12:21 +0000)]
avcodec/decode: reset codec on receiving packet after EOF in compat_decode

In commit 061a0c14bb57 ("decode: restructure the core decoding code"), the
deprecated avcodec_decode_* APIs were reworked so that they called into the
new avcodec_send_packet / avcodec_receive_frame API. This had the side effect
of prohibiting sending new packets containing data after a drain
packet, but in previous versions of FFmpeg this "worked" and some
applications relied on it.

To restore some compatibility, reset the codec if we receive a new non-drain
packet using the old API after draining has completed. While this does
not give the same behaviour as the old API did, in the majority of cases
it works and it does not require changes to any other part of the decoding
code.

Fixes ticket #6775
Signed-off-by: James Cowgill <jcowgill@debian.org>
Signed-off-by: Marton Balint <cus@passwd.hu>
23 months agoarm/hevc_idct: fix compilation on Android
James Almer [Fri, 8 Dec 2017 22:14:13 +0000 (19:14 -0300)]
arm/hevc_idct: fix compilation on Android

Compilation error "out of range" fixed for armeabi-v7a. Compilation failed
trying to build libvlc.aar for ARM7 android on ubuntu 16.04 host. Error
messages is "Offset out of range". The reason of the error is assembler LDR
directives in function "ff_hevc_transform_luma_4x4_neon_8" need local storage
in range <1k, but no such storage provided.

Based on a patch by Ihor Bobalo <bob@eleks.com>

Suggested-by: wbs
Signed-off-by: James Almer <jamrial@gmail.com>
23 months agocheckasm/llviddsp : add test for add_gradient_pred
Martin Vignali [Sat, 2 Dec 2017 18:47:24 +0000 (19:47 +0100)]
checkasm/llviddsp : add test for add_gradient_pred

23 months agoavcodec/utvideodec : add SIMD (SSSE3 and AVX2) for gradient_pred
Martin Vignali [Sat, 2 Dec 2017 18:46:42 +0000 (19:46 +0100)]
avcodec/utvideodec : add SIMD (SSSE3 and AVX2) for gradient_pred

23 months agoavcodec/x86/lossless_videodsp : add avx2 version for add_left_pred
Martin Vignali [Sat, 2 Dec 2017 18:09:58 +0000 (19:09 +0100)]
avcodec/x86/lossless_videodsp : add avx2 version for add_left_pred

23 months agoavcodec/x86/lossless_videodsp.asm : make macro for add_left_pred_unaligned in order...
Martin Vignali [Sat, 2 Dec 2017 18:04:11 +0000 (19:04 +0100)]
avcodec/x86/lossless_videodsp.asm : make macro for add_left_pred_unaligned in order to add avx2 version

23 months agocheckasm/llviddsp : test return of add_left_pred(16)
Martin Vignali [Sat, 2 Dec 2017 18:03:30 +0000 (19:03 +0100)]
checkasm/llviddsp : test return of add_left_pred(16)

23 months agoavcodec/huffyuvenc : sub_left_prediction_bgr32, call dsp after 32 first byte of the...
Martin Vignali [Sat, 2 Dec 2017 17:33:25 +0000 (18:33 +0100)]
avcodec/huffyuvenc : sub_left_prediction_bgr32, call dsp after 32 first byte of the line in order to try to call the align version (diff bytes have avx2 version)

23 months agoavcodec/huffyuvenc : remove code duplication in sub_left_prediction
Martin Vignali [Sat, 2 Dec 2017 17:32:05 +0000 (18:32 +0100)]
avcodec/huffyuvenc : remove code duplication in sub_left_prediction

start of the line (before dsp call), can be merge with width < 32 part

23 months agoavcodec/huffyuvenc : increase scalar loop count
Martin Vignali [Sat, 2 Dec 2017 17:30:26 +0000 (18:30 +0100)]
avcodec/huffyuvenc : increase scalar loop count

in order to try to call dsp in aligned mode
(diff_int16 have AVX2 now)

23 months agoavcodec/utvideodec : use dsp add_median_pred for second line
Martin Vignali [Sun, 26 Nov 2017 18:18:48 +0000 (19:18 +0100)]
avcodec/utvideodec : use dsp add_median_pred for second line

process start of the line in scalar, before call dsp
(dsp need align 16)

23 months agocheckasm/vf_threshold : add test for threshold16
Martin Vignali [Thu, 7 Dec 2017 21:03:44 +0000 (22:03 +0100)]
checkasm/vf_threshold : add test for threshold16

23 months agoavfilter/x86/vf_threshold : add threshold16 SIMD (SSE4 and AVX2)
Martin Vignali [Thu, 7 Dec 2017 21:01:54 +0000 (22:01 +0100)]
avfilter/x86/vf_threshold : add threshold16 SIMD (SSE4 and AVX2)

23 months agohevc: Add hevc_get_pixel_4/8/12/16/24/32/48/64
Alexandra Hájková [Sun, 3 Dec 2017 08:15:23 +0000 (09:15 +0100)]
hevc: Add hevc_get_pixel_4/8/12/16/24/32/48/64

Checkasm timings:
block size bitdepth  C       NEON
4           8 bit:    146.7   48.7
           10 bit:    146.7   52.7
8           8 bit:    430.3   84.4
           10 bit:    430.4  119.5
12          8 bit:    812.8  141.0
           10 bit:    812.8  195.0
16          8 bit:   1499.1  268.0
           10 bit:   1498.9  368.4
24          8 bit:   4394.2  574.8
           10 bit:   3696.3  804.8
32          8 bit:   5108.6  568.9
           10 bit:   4249.6  918.8
48          8 bit:  16819.6 2304.9
           10 bit:  13882.0 3178.5
64          8 bit:  13490.8 1799.5
           10 bit:  11018.5 2519.4

Signed-off-by: Martin Storsjö <martin@martin.st>
23 months agoavfilter/vf_scale: add more aliases for "range" options
Paul B Mahol [Fri, 8 Dec 2017 19:35:08 +0000 (20:35 +0100)]
avfilter/vf_scale: add more aliases for "range" options

Signed-off-by: Paul B Mahol <onemda@gmail.com>
23 months agofate: Fix fate-mov-bbi-elst-starts-b on ARM
Michael Niedermayer [Fri, 8 Dec 2017 15:58:10 +0000 (16:58 +0100)]
fate: Fix fate-mov-bbi-elst-starts-b on ARM

23 months agofate/mxf: add reel name test
Mark Reid [Tue, 5 Dec 2017 04:46:23 +0000 (20:46 -0800)]
fate/mxf: add reel name test

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavformat/mxfenc: write reel_name if metadata key is present
Mark Reid [Tue, 5 Dec 2017 04:46:22 +0000 (20:46 -0800)]
avformat/mxfenc: write reel_name if metadata key is present

Reviewed-by: Tomas Härdin <tjoppen@acc.umu.se>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavformat/mxfenc: use track count to generate component instance uuid
Mark Reid [Tue, 5 Dec 2017 04:46:21 +0000 (20:46 -0800)]
avformat/mxfenc: use track count to generate component instance uuid

Reviewed-by: Tomas Härdin <tjoppen@acc.umu.se>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agolibavcodec/hevc_filter: move AVDISCARD_NONREF switch-case into function
sfan5 [Thu, 7 Dec 2017 19:37:48 +0000 (20:37 +0100)]
libavcodec/hevc_filter: move AVDISCARD_NONREF switch-case into function

In preparation for implementation of skip_frame.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavcodec/nvdec: Fix capability check with old drivers.
Jacob Trimble [Thu, 30 Nov 2017 20:35:50 +0000 (12:35 -0800)]
avcodec/nvdec: Fix capability check with old drivers.

Copied the check from cuviddec.c (*_cuvid decoders) to allow the
capability check to be optional for older drivers.

Signed-off-by: Jacob Trimble <modmaker@google.com>
Signed-off-by: Timo Rothenpieler <timo@rothenpieler.org>
23 months agovf_zscale: Fix alpha destination graph for floating point pixel formats
Vittorio Giovara [Thu, 7 Dec 2017 21:32:22 +0000 (16:32 -0500)]
vf_zscale: Fix alpha destination graph for floating point pixel formats

This was setting the input pixel type instead of the output one,
leading to incorrect data being fed to the library.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
23 months agoavformat: add NSP demuxer
Paul B Mahol [Fri, 1 Dec 2017 14:15:13 +0000 (15:15 +0100)]
avformat: add NSP demuxer

Signed-off-by: Paul B Mahol <onemda@gmail.com>
23 months agoavfilter/vf_waveform: add default case when picking input formats
Paul B Mahol [Fri, 8 Dec 2017 11:31:01 +0000 (12:31 +0100)]
avfilter/vf_waveform: add default case when picking input formats

Should silence compiler warnings.

Signed-off-by: Paul B Mahol <onemda@gmail.com>
23 months agoopus: fix hybrid folding indexing during band quantization
Rostislav Pehlivanov [Fri, 8 Dec 2017 03:40:41 +0000 (03:40 +0000)]
opus: fix hybrid folding indexing during band quantization

Resulted in valgrind errors due to uninitialized memory.
Also updates fate and makes it use the tron sample result.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
23 months agofate-opus: run and test inactive samples
Rostislav Pehlivanov [Tue, 5 Dec 2017 02:03:15 +0000 (02:03 +0000)]
fate-opus: run and test inactive samples

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>
23 months agox86/vf_hflip: use xor to zero initialize registers
James Almer [Thu, 7 Dec 2017 22:28:19 +0000 (19:28 -0300)]
x86/vf_hflip: use xor to zero initialize registers

Signed-off-by: James Almer <jamrial@gmail.com>
23 months agox86/vf_hflip: don't load the width argument twice
James Almer [Thu, 7 Dec 2017 21:09:00 +0000 (18:09 -0300)]
x86/vf_hflip: don't load the width argument twice

Signed-off-by: James Almer <jamrial@gmail.com>
23 months agolavc/libx265: mark disposable frames
John Stebbins [Wed, 6 Dec 2017 00:36:58 +0000 (16:36 -0800)]
lavc/libx265: mark disposable frames

Used by movenc to fill sdtp box

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agovaapi_h265: Add named options for setting profile and level
Mark Thompson [Wed, 29 Nov 2017 21:31:53 +0000 (21:31 +0000)]
vaapi_h265: Add named options for setting profile and level

Also fixes the default, which previously contained a nonsense value.

23 months agovaapi_h264: Add named options for setting profile and level
Mark Thompson [Wed, 29 Nov 2017 21:31:11 +0000 (21:31 +0000)]
vaapi_h264: Add named options for setting profile and level

23 months agoavcodec/libx265 - Add named option to set profile
Gyan Doshi [Tue, 5 Dec 2017 07:47:53 +0000 (13:17 +0530)]
avcodec/libx265 - Add named option to set profile

Adds call to x265_param_apply_profile after x265_param_parse.
Added as private option since HEVC profiles other than
Main, Main 10 and MSP in AVCodecContext are consolidated in a single
constant.

Reviewed-by: Hendrik Leppkes <h.leppkes@gmail.com>
Reviewed-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Reviewed-by: Lou Logan <lou@lrcd.com>
23 months agoavcodec/diracdsp: Fix integer overflow in PUT_SIGNED_RECT_CLAMPED()
Michael Niedermayer [Sat, 2 Dec 2017 20:53:22 +0000 (21:53 +0100)]
avcodec/diracdsp: Fix integer overflow in PUT_SIGNED_RECT_CLAMPED()

Fixes: runtime error: signed integer overflow: 2147483646 + 2048 cannot be represented in type 'int'
Fixes: 4479/clusterfuzz-testcase-minimized-6529894147162112

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavcodec/dirac_dwt: Fix integer overflows in COMPOSE_DAUB97*
Michael Niedermayer [Sat, 2 Dec 2017 20:48:04 +0000 (21:48 +0100)]
avcodec/dirac_dwt: Fix integer overflows in COMPOSE_DAUB97*

Fixes: 4478/clusterfuzz-testcase-minimized-4752113767809024
Fixes: runtime error: signed integer overflow: -2147483626 + -319489 cannot be represented in type 'int'

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavformat/mux: stop delaying writing the header
James Almer [Sun, 26 Nov 2017 19:47:26 +0000 (16:47 -0300)]
avformat/mux: stop delaying writing the header

Every bitstream filter behaves as intended now, so there's no need to
wait for the first packet of every stream.

Signed-off-by: James Almer <jamrial@gmail.com>
23 months agoavcodec/extract_extradata_bsf: Fix leak discovered via fuzzing
Nikolas Bowe [Tue, 5 Dec 2017 23:11:26 +0000 (15:11 -0800)]
avcodec/extract_extradata_bsf: Fix leak discovered via fuzzing

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
23 months agoavfilter: add setrange filter
Paul B Mahol [Sun, 3 Dec 2017 10:59:58 +0000 (11:59 +0100)]
avfilter: add setrange filter

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agotests/fate/mov: Fix fate-mov-invalid-elst-entry-count failure on ARM
Michael Niedermayer [Wed, 6 Dec 2017 00:03:28 +0000 (01:03 +0100)]
tests/fate/mov: Fix fate-mov-invalid-elst-entry-count failure on ARM

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agofate/hevc: add skip_loop_filter test
sfan5 [Tue, 5 Dec 2017 11:47:47 +0000 (12:47 +0100)]
fate/hevc: add skip_loop_filter test

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2 years agodoc/developer: revise mailing list section
Jim DeLaHunt [Mon, 4 Dec 2017 07:09:02 +0000 (23:09 -0800)]
doc/developer: revise mailing list section

The Developer Documentation had instructions to
subscribe to the ffmpeg-cvslog email list. But that is
no longer accurate. For the purposes in this section --
review of patches, discussion of development issues --
ffmpeg_devel is the appropriate email list. Some developers
may want to monitor ffmpeg-cvslog, but it is not mandatory.

This is v3 of this doc, based on discussion in thread
<https://ffmpeg.org/pipermail/ffmpeg-devel/2017-November/220528.html>
and in response to docs Maintainer comments in
<https://ffmpeg.org/pipermail/ffmpeg-devel/2017-December/221596.html>.

1. In doc/developer.texi, add a new section about
ffmpeg-devel, based on existing text from ffmpeg-cvslog
section regarding discussion of patches and of
development issues. Reflect wording from discussion at
<https://ffmpeg.org/pipermail/ffmpeg-devel/2017-November/221199.html>
but with copy-editing to make wording more concise.

2. In doc/developer.texi, rewrite the ffmpeg-cvslog section
to match the current usage of ffmpeg-cvslog. Some
developers choose to follow this list, but it is not
mandatory.

There are a lot of improvements possible to the
Developer Documentation page, beyond this refactoring.
However, making those improvements is a much bigger
and more difficult task.  This change is "low hanging
fruit".

Signed-off-by: Jim DeLaHunt <from.ffmpeg-dev@jdlh.com>
Signed-off-by: Timothy Gu <timothygu99@gmail.com>
2 years agovaapi_h265: Add named options for setting profile and level
Mark Thompson [Wed, 29 Nov 2017 21:31:53 +0000 (21:31 +0000)]
vaapi_h265: Add named options for setting profile and level

Also fixes the default, which previously contained a nonsense value.

2 years agovaapi_h264: Add named options for setting profile and level
Mark Thompson [Wed, 29 Nov 2017 21:31:11 +0000 (21:31 +0000)]
vaapi_h264: Add named options for setting profile and level

2 years agorkmpp: Add hardware config information
Mark Thompson [Fri, 1 Dec 2017 22:01:41 +0000 (22:01 +0000)]
rkmpp: Add hardware config information

This is not strictly required here because the rkmpp decoder does not
call ff_get_format(), but it may be helpful metadata for users.

2 years agoavfilter/vf_convolution: add 7x7 filter
Paul B Mahol [Mon, 4 Dec 2017 20:32:29 +0000 (21:32 +0100)]
avfilter/vf_convolution: add 7x7 filter

Signed-off-by: Paul B Mahol <onemda@gmail.com>
2 years agox86/vf_threshold: make threshold8 functions work on x86_32
James Almer [Mon, 4 Dec 2017 18:04:43 +0000 (15:04 -0300)]
x86/vf_threshold: make threshold8 functions work on x86_32

Signed-off-by: James Almer <jamrial@gmail.com>