Merge commit 'cd4d9df22738e6f147521ccb72c7930db6050914'

* commit 'cd4d9df22738e6f147521ccb72c7930db6050914':
  asfdec: free AVDictionaries properly when closing the demuxer

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

Merge commit '90e648c508fa6b97587ce62699a1449123f7d285'

* commit '90e648c508fa6b97587ce62699a1449123f7d285':
  fate/mp3: use the f32le format as output

Merged-by: Hendrik Leppkes <h.leppkes@gmail.com>

avcodec/hevc_mp4toannexb_bsf: Remove ;;

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/mov: fix regression in processing .aax files

Commit 0a551cbe introduced "activation_bytes" option, and not specifying
this option (while calling ffmpeg / ffprobe) causes the program to quit
early. Before this commit, ffprobe was capable of processing metadata in
.aax files.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avfilter/avf_showvolume: stop making output fully transparent

aacenc: add description to the 'aac_coder' option

This commit adds a short description to the aac_coder option of the
AAC encoder in order to be consistent with the other options.
Generally, right now, the 'FAAC' method works fine with speech and
low broadband spectrum audio. 'Fast' is just as the name suggests.
'ANMR' still needs work and 'Twoloop', the default, works well with
every type of audio.

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>

Revert "libavcodec/utils: av_get_exact_bits_per_sample fixed to return "1" for DSD codecs."

Requested-by: Ganesh Ajjanagadde <gajjanag@mit.edu> (original patch reviewer)
This reverts commit 293e25be615d43c9b611c4b04920ec0d245a7145.

Merge branch 'simple_http_server_api_implementation' of http://git.klaxa.eu/git/ffmpeg

Reviewed-by: Nicolas George <george@nsup.org>
Merged-by: Michael Niedermayer <michael@niedermayer.cc>

aacenc: remove redundant argument from coder functions

This commit removes a redundant argument from the functions in aaccoder.
The argument lambda was redundant as it was just a copy of s->lambda,
to which all functions have access to anyway. This cleans up the function
pointers a bit which is helpful as there are a lot of other search_for_*
functions under development and with them populated it gets messy.

Reviewed-by: Claudio Freire <klaussfreire@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/libx264: expose nv21 input support

libx264 added support for NV21 input recently.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

libavcodec/utils: av_get_exact_bits_per_sample fixed to return "1" for DSD codecs.

All DSD codes have 1 bit per sample.
https://en.wikipedia.org/wiki/Direct_Stream_Digital

Signed-off-by: Ihar A. Tumashyk <itumashyk@gmail.com>
Reviewed-by: Ganesh Ajjanagadde <gajjanag@mit.edu>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

doc/example: Add http multi-client example code

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

doc/protocols: document experimental mutli-client api

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/http: Implement server side network code.

add http_accept,
add http_handshake and move handshake logic there,
handle connection closing.

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/tcp: increase range for listen and call the underlying socket operations accordingly

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/tcp: add tcp_accept

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/avio: add avio_accept and avio_handshake

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/avio: add ffurl_accept and ffurl_handshake

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/network: split ff_listen_bind into ff_listen and ff_accept

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

MAINTAINERS: update my maintership

dshow: show more debug timestamp info

Signed-off-by: rogerdpack <rogerpack2005@gmail.com>

MAINTAINERS: Remove myself as leader

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit 'a0797950527120c85263c910eb6ba08fddcfdcb3'

* commit 'a0797950527120c85263c910eb6ba08fddcfdcb3':
  fate/mp3: specify the number of output samples instead of filesize

Merged-by: Michael Niedermayer <michael@niedermayer.cc>

asfdec: free AVDictionaries properly when closing the demuxer

Signed-off-by: Anton Khirnov <anton@khirnov.net>

fate/mp3: use the f32le format as output

Those tests test the float decoder, so it is preferable to dump the
decoder output directly instead of converting it to s16.

fate/mp3: specify the number of output samples instead of filesize

This is not dependent on the output format.

avformat/wavdec: add extra sample count check for G.729 files

Can be used to fix Ticket4577

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

mips/hevcdsp: fix string concatenation on macros

Needed for old compilers like GCC 4.2

Tested by trac user brad. Fixes ticket #4745

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>

x86/aacpsdsp: add SSE and SSE3 optimized functions

Between 1.5 and 2.5 times faster

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>

Revert "ffmpeg: modify tty state when stderr is redirected"

faults in fate otherwise breaks the terminal.
To reproduce, add a abort() into wav_read_header()
run make fate-acodec-adpcm-ima_wav

This reverts commit 92e62f49cf7440a9e8998d284528e223c0948c97.

avcodec/r210enc: Drop coded_frame usage

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/avuienc: Drop coded_frame usage

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit '764ec70149728be82304c163ccc4e280f1629201'

* commit '764ec70149728be82304c163ccc4e280f1629201':
  rtsp: Only interpret $ as interleaved packet indicator at the start of replies

Merged-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit '6877537ff321136cf78ab03b8b91e6b6f218db5b'

* commit '6877537ff321136cf78ab03b8b91e6b6f218db5b':
  hmac: add missing version bump and APIChanges entry

Conflicts:
libavutil/version.h

version.h is not changed as the code has been added 2 years ago in FFmpeg

See: 82ef67016ef731268ab7b8e91139bd5f16edc991
Merged-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit '0c73a5a53cc97f4291bbe9e1e68226edf6161744'

* commit '0c73a5a53cc97f4291bbe9e1e68226edf6161744':
  hls: Save and forward avio options

Conflicts:
libavformat/hls.c

Merged-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit '5ae178539b91d25710b7bb322d156c31aea9f8bf'

* commit '5ae178539b91d25710b7bb322d156c31aea9f8bf':
  http: Add the trailing endlines if they are missing

Merged-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/y41penc: Drop coded_frame usage

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/yuv4enc: drop coded_frame usage

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

rtsp: Only interpret $ as interleaved packet indicator at the start of replies

Allow $ as character anywhere within normal RTSP replies - both
within the lines, and as the first character of RTSP header lines.
(The existing old comment indicated that an inline packet could
start at any line within a RTSP reply header, but that doesn't
sound valid to me, and I'm not sure if the existing code
handled that correctly either.)

CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>

avcodec/v308enc: Remove coded_frame usage

coded_frame is deprecated

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/v408enc: remove coded_frame usage

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

hmac: add missing version bump and APIChanges entry

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Martin Storsjö <martin@martin.st>

ffmpeg: modify tty state when stderr is redirected

This fixes Ticket2964

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit 'bf0cef5c3a114df452e5476167634dd8f51eb448'

* commit 'bf0cef5c3a114df452e5476167634dd8f51eb448':
  checkasm: Include io.h for isatty, if available

Merged-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit '9487ffd4c02b9e261562d43735490068c0df0d4b'

* commit '9487ffd4c02b9e261562d43735490068c0df0d4b':
  rtmpproto: free hmac context properly

See: 6ce3d9fc4a70958680b93c6995cc87f34630a7e3
Merged-by: Michael Niedermayer <michael@niedermayer.cc>

hls: Save and forward avio options

Make possible to send the custom headers and override the user agent.

Reported-by: BenWonder

http: Add the trailing endlines if they are missing

Makes slightly easier the life of those want to use the option
from the command line

checkasm: Include io.h for isatty, if available

configure does check for isatty, and checkasm properly checks
HAVE_ISATTY, but on some platforms (e.g. WinRT), io.h needs to be
included for isatty to be available.

Signed-off-by: Martin Storsjö <martin@martin.st>

rtmpproto: free hmac context properly

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Martin Storsjö <martin@martin.st>

avcodec/mpegvideo_enc: Ignore QMAX if VBV constraints are exceeded

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/mpeg12enc: extend QP range to 28 for non linear quantizers

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

libvpxenc: report pict_type/error via side-data

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Zern <jzern@google.com>

libvpxenc: quiet coded_frame deprecation warnings

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Zern <jzern@google.com>

avformat/rtmpproto: free hmac context properly

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

avfilter/vf_pnsr: fix author name

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

doc/APIchanges: fill in some missing things

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avutil/hmac: fix API/ABI compatibility with the fork

The test code is updated with some cosmetics to avoid the loop
using undefined AVHMACType values.
The old enum values will be removed in the next major bump, effectively
making both projects synced and without API or ABI issues.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>

Merge commit '65dd6a1f84cc5c9a91d279e8e8b72ff7adf58d31'

* commit '65dd6a1f84cc5c9a91d279e8e8b72ff7adf58d31':
  rtmpproto: use AVHMAC instead of a custom implementation

See: 6c87b866285fae0a45ab4d7162088428f60bd332
Merged-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit 'e59f7cd89e7f4f096f8456d63191adec2f8b1083'

* commit 'e59f7cd89e7f4f096f8456d63191adec2f8b1083':
  fate: add test vectors for HMAC SHA and SHA-2

Conflicts:
libavutil/hmac.c
tests/ref/fate/hmac

See: 1163910a00595f01f5f6aabd5d69b01925429401
Merged-by: Michael Niedermayer <michael@niedermayer.cc>

avutil/hmac: support another ABI difference to the fork

spotted this and so adding it.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit '7e38340b831af8e3271f22bd66b9da14953c02af'

* commit '7e38340b831af8e3271f22bd66b9da14953c02af':
  hmac: add support for SHA-2

Conflicts:
libavutil/hmac.c
libavutil/hmac.h

See: 82ef67016ef731268ab7b8e91139bd5f16edc991
Merged-by: Michael Niedermayer <michael@niedermayer.cc>

rawdec: fix mjpeg probing buffer size check

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

rawdec: fix mjpeg probing

There can be other headers than "Content-Type:" (in this case, a
"Content-Length:" header was following), so checking for a trailing
newline is wrong.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/oggdec: ogg_read_seek: reset ogg after seeking

Fixes Ticket4743

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

rtmpproto: use AVHMAC instead of a custom implementation

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Martin Storsjö <martin@martin.st>

fate: add test vectors for HMAC SHA and SHA-2

Also replace custom tests for MD5 with those published in RFC 2202

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Martin Storsjö <martin@martin.st>

hmac: add support for SHA-2

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Martin Storsjö <martin@martin.st>

avformat/rtmpproto: use AVHMAC instead of a custom implementation

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: James Almer <jamrial@gmail.com>

ffmpeg_op: Print warning if duration isnt known when -sseof is used

Found-by: ubitux
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

doc/ffmpeg: Use @code

Found-by: ubitux
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/dvbsubdec: Implement display definition segment fallback from  ETSI EN 300 743 V1.3.1

Fixes: subtitles.ts
Fixes: Ticket679
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffmpeg: Implement support for seeking relative to EOF

Fixes Ticket227

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/aacdec_fixed: Fix preparation for resampler

2nd channel makes sense only for CPE type.

Skip 2nd channel in preparation for resampler (in spectral_to _sample())
depending on block type.

Fixes fate failure with clang ftrapv.

Signed-off-by: Nedeljko Babic <nedeljko.babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffplay: Use sws_scale to scale subtitles

Fixes some files from Ticket679

This also changes subtitles to 4:2:0 matching the output format and thus
simplifying the blend code.
This restricts placement to the chroma sample resolution though, speak up
if you consider this a problem, say so, the code could be changed to use
YUV444 for subtitles and scaling them down while blending, this would be
slower though.
The current code only uses a single swscale context and reinitializes it
as needed, this could be changed as well if needed

Reviewed-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/wavdec: Check for data_size overflow

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

wavdec: make sample count check more precise

May be used to fix Ticket4577

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/h264_refs: Allow slightly larger pps_ref_count[0] in PAFF RAP detection heuristic

Fixes Ticket4738

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/h264_refs: extend RAP heuristic to multiple PPS

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/utils: Set coded_frame.pict_type in generic code

This makes it possible to remove more coded_frame usage without breaking the
publically visible coded_frame

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/proresenc_anatoliy: remove coded_frame use

coded_frame.key_frame should be set by the core already

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/avcodec: define CODEC_FLAG based on AV_CODEC_FLAG instead of duplicating the numbers and comments

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit 'c6e0829e3696c894d24c8088c3ee30a9c17534c2'

* commit 'c6e0829e3696c894d24c8088c3ee30a9c17534c2':
  configure: Don't force _WIN32_WINNT to an older version if targeting winphone/winrt

Merged-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit 'd75b55635a02444c2f188c26e431a1cec992babe'

* commit 'd75b55635a02444c2f188c26e431a1cec992babe':
  dxva2/d3d11va: Set _WIN32_WINNT to 0x0602 instead of 0x0600

Merged-by: Michael Niedermayer <michael@niedermayer.cc>

libavcodec/qsvdec.c delay in 1 microsecond replaced to more appropriate 500 microseconds

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

libavcodec/qsvenc.c delay in 1 microsecond replaced to more appropriate 500 microseconds

This commit replaces the 1 microsecond delay by 500 microsecond for the
case when the MFX library does return MFX_WRN_DEVICE_BUSY status.
In general this warning never appears for simple encoding or
transcoding session because the GPU is so fast so it almost always is not busy and
any delay value just does not executes.
But for heavy transcoding tasks for example, when several QSV sessions
are running simultaneously then using a 1-microsecond delay may
result in 1000 iterations per each frame.
So here possible a paradoxical case when GPU loading also loads CPU by dummy tasks.
Official MFX/QSV samples by Intel are using 1 millisecond (i.e. 1000
microseconds) everywhere where MFX_WRN_DEVICE_BUSY does appear.
So 500us is a much more optimal value than 1us.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/aac_fixed: Fix a bug in spectral_to_sample()

There was fixed number of loops (2048) in preparation for resampler, so
when number of samples is smaller than this, there would be an overflow on
ret_buf.

For some reason this behavior popped out only under valgrind with
--disable-memory-poisoning option.

This is now fixed and number of loops depends on actual number of samples.

Signed-off-by: Nedeljko Babic <nedeljko.babic@rt-rk.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffmpeg.c: remove all remaining coded_frame uses

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

configure: Don't force _WIN32_WINNT to an older version if targeting winphone/winrt

This avoids having to manually set _WIN32_WINNT in --extra-cflags
when targeting these API families, which only was necessary to
work around configure setting _WIN32_WINNT to an older version
by default.

Signed-off-by: Martin Storsjö <martin@martin.st>

dxva2/d3d11va: Set _WIN32_WINNT to 0x0602 instead of 0x0600

If _WIN32_WINNT is unset, we force it to a new enough value to
make sure the necessary definitions are visible.

When targeting Windows Phone or Windows RT, _WIN32_WINNT should
be at least 0x0602 - otherwise the windows headers themselves
can cause errors (which technically are bugs in the headers).

Raising this value here shouldn't hurt; the alternative would
be to not touch it at all if WINAPI_FAMILY is set to phone/app,
or to force setting it to 0x0602 in configure if unset (for phone/app).

Signed-off-by: Martin Storsjö <martin@martin.st>

libavcodec/hevc_mp4toannexb_bsf.c: Optional argument "private_spspps_buf" to avoid extradata modification.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/libdcadec: exss is used only under ifdef, thus mark it as potentially unused

avoids a compiler warning

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/h264_slice: Also check sei_recovery_frame_cnt for skip_frame nokey

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/h264_slice: do not skip 2nd field if first was not skipped

Found-by: John Högberg <john.hogberg@ericsson.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avdevice/decklink: Fix build error caused by a change in the SDK.

In version 10.4 of the DeckLink SDK, GetBufferedAudioSampleFrameCount() was changed to take an unsigned int instead of an unsigned long.

Signed-off-by: Chris Spencer <spencercw@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

ffmpeg: Use side data instead of coded_frame for error[] values

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/snowenc: Export quality/pict_type/PSNR stats through side data

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/mpegvideo_enc: export per frame PSNR through side data

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avocdec/mpeg12dec: fix mpeg_xvmc hwaccel flag

The flag is deprecated and as such didn't get a new AV_ prefixed version

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

avcodec/mpegvideo_enc: move ff_side_data_set_encoder_stats() call to after the vbv retry code

This ensures the data matches the final values

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Merge commit 'b7040e67ec18259ca634a0e29d98469b3484a87c'

* commit 'b7040e67ec18259ca634a0e29d98469b3484a87c':
  h264: fix AVDISCARD_NONKEY for some interlaced content

Conflicts:
libavcodec/h264.c

Not merged, the used field is not initialized where it is used

Merged-by: Michael Niedermayer <michael@niedermayer.cc>