Merge commit '75439bb31b3ab76f9a4d30c5de4ff9f407ec8128'

* commit '75439bb31b3ab76f9a4d30c5de4ff9f407ec8128':
  siff: K&R formatting cosmetics

Conflicts:
libavformat/siff.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

avformat/img2dec: remove the non-negative number requirement for start_number option

industrial cameras usually mark the trigger frame as frame number 0
all frames saved before trigger frame receive a negative sequence number

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/tta: fix crash with corrupted files

av_add_index_entry() can fail, for example because the parameters are
invalid, or because memory allocation fails. Check this; it can actually
happen with corrupted files.

The second hunk is just for robustness. Just in case functions like
ff_reduce_index() remove entries. (Not sure if this can actually
happen.)

Fixes ticket #4294.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

siff: K&R formatting cosmetics

avformat/mpegtsenc: Fix build with msvc

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avutil/mem: replace remaining void **/*** casts by memcpy()

This is similar to 60392480181f24ebf3ab48d8ac3614705de90152

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/mov: simplify pasp value before setting SAR

This avoids a 65536:65536 SAR

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/ppc/idctdsp.c: POWER LE support in idct_add_altivec()

also add GET_TMP2() macro

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/network: Check for av_malloc* failures in ff_tls_init()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

swresample: Use int instead of enum for fields which are accessed through AVOptions as int

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Revert "avformat/mpegtsenc: Change the service_type field to enum"

This isnt safe due to AVOption accessing this field as int
i just realized a moment after pushing

This reverts commit 0935453e84a17035f0c3544913e7ab7775839176.

avformat/mpegtsenc: Change the service_type field to enum

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavformat/mpegtsenc: allow to set service_type in sdt

This adds an option to set the service type in mpegts as defined in ETSI 300 468.

I added what I believe are the most useful service types as pre defined values,
the others can be sent by using their hexdecimal form directly (e.g. -mpegts_service_type digital_radio, -mpegts_service_type 0x07).

I've been using this patch in order to pipe internet radio stream (originally as HLS/m3u8) from ffmpeg to tvheadend,
when the service type set right tvheadend recognize the mpegts stream as a radio channel.

The patch in its original form was written by linuxstb from freenode's hts channel which allowed me pushing it upstream.

This close issue 4118.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '11e05533170485b593974cf90916425a0188e7bd'

* commit '11e05533170485b593974cf90916425a0188e7bd':
  Ignore generated file tools/sidxindex.

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '3d5d46233cd81f78138a6d7418d480af04d3f6c8'

* commit '3d5d46233cd81f78138a6d7418d480af04d3f6c8':
  opus: Factor out imdct15 into a standalone component

Conflicts:
configure
libavcodec/opus_celt.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '28df0151b6618226b05ee52e031af0b11ca531b0'

* commit '28df0151b6618226b05ee52e031af0b11ca531b0':
  configure: Add a dependency on vc1_decoder from vc1_parser

See: 6ac3c8c6a0c6c69b84e408a9b494fcb26294b9ee
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Ignore generated file tools/sidxindex.

avcodec/hevc_filter: Fix indention

Found-by: Timothy Gu <timothygu99@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/hevc: adding support for monochrome sequences in hevc

cherry picked from commit 8e50557707d2ec11ccad657470b2e140f314348e

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc/libopenjpegenc: move opj_create_compress, opj_cio_open and opj_set_event_mgr to libopenjpeg_encode_frame

 libopenjpegenc crashes with "pointer being freed was not allocated" when threading
 is enabled with:
 ffmpeg -i tests/vsynth1/01.pgm -vcodec libopenjpeg file.j2k
 this appears to be a bug in libopenjpeg

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc/libopenjpegenc: factorize cinema parameters to it's own function

Signed-off-by: Jean First <jeanfirst@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

opus: Factor out imdct15 into a standalone component

It will be reused by the AAC decoder.

configure: Add a dependency on vc1_decoder from vc1_parser

This fixes builds with vc1_parser enabled without vc1_decoder. All
the vc1_decoder object files were included in the vc1_parser line
in libavcodec/Makefile before, but architecture specific object files
for vc1_decoder were not.

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

avcodec/wavpackenc: remove unneeded L suffixes

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

doc/ffprobe.xsd: drop build_date and build_time from programVersionType

Update schema after commit:
103e4c58633f210d7c3203cae121ba5b5123afa6

doc/doxy-wrapper.sh: autodetect version

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

stop embedding the build date

Theis makes the build binary reproducible.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

x86/hevcdsp: add missing vzeroupper in ff_hevc_sao_band_filter_48_*_avx2

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

vorbis: parse out setup headers as well

Prevents an 'Invalid packet' message. Currently mid-stream setup packets
are ignored. Theoretically, they could, based on the specification, be used to
reinitialize the stream if parameters change, but I don't expect that to be
common (and no one seems to have asked for it).

Signed-off-by: Ben Boeckel <mathstuf@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffplay: Fallback to dts if pts is unavailable in pkt_in_play_range calculation

Reviewed-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

x86/hevcdsp: add missing guards to ff_hevc_sao_band_filter_avx2

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

lavu/x86/x86inc: deprecate INIT_AVX

The same can be done with INIT_XMM avx

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

hevc/sao: use aligned copies

For band filter, source and destination are aligned (except for 16x16 ctbs),
and otherwise, they are most often aligned. Overall, the total width is also
too small for amortizing memcpy.

Timings (using an intrinsic version of edge filters):
          B/32     B/64     E/32     E/64
Before:  32045    93952    38925    126896
After:   26772    83803    33942    117182

x86: hevc/sao: aligned source buffers

Usefull for at least band filter, for which:
- Band filter call only:
           32      64
Before:  16556    54015
After:   16497    52355
- Whole case:
           32      64
Before:  37031   103008
After:   32045    93952

x86/hevc: add ff_hevc_sao_band_filter_{8,10,12}_{sse2,avx,avx2}

Original x86 intrinsics code and initial 8bit yasm port by Pierre-Edouard Lepere.
10/12bit yasm ports, refactoring and optimizations by James Almer

Benchmarks of BQTerrace_1920x1080_60_qp22.bin with an Intel Core i5-4200U

width 32
40338 decicycles in sao_band_filter_0_8, 2048 runs, 0 skips
8056 decicycles in ff_hevc_sao_band_filter_8_32_sse2, 2048 runs, 0 skips
7458 decicycles in ff_hevc_sao_band_filter_8_32_avx, 2048 runs, 0 skips
4504 decicycles in ff_hevc_sao_band_filter_8_32_avx2, 2048 runs, 0 skips

width 64
136046 decicycles in sao_band_filter_0_8, 16384 runs, 0 skips
28576 decicycles in ff_hevc_sao_band_filter_8_32_sse2, 16384 runs, 0 skips
26707 decicycles in ff_hevc_sao_band_filter_8_32_avx, 16384 runs, 0 skips
14387 decicycles in ff_hevc_sao_band_filter_8_32_avx2, 16384 runs, 0 skips

Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>

x264asm: warn when inappropriate instruction used in function with specified cpuflags

Requested-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Requested-by: "Ronald S. Bultje" <rsbultje@gmail.com>

tests/utils: Remove unneeded L suffix

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mpegvideo_enc: Fix number suffixes in rc_buffer_size calculation

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/h264_cabac: use int instead of long for mbb_xy

The mb address fits in int

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/omadec: fix number suffix

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/smacker: Fix number suffix

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/matroskadec: Fix number suffixes

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/dxtory: Use LL instead of L number suffix

This is probably unneeded and normal int would be fine, but its
safer to use LL and this isnt speed relevant

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

swresample/dither: Cleanup number suffixes

The <<31 case needs LL

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

hevcdsp: replace the SAOParams struct parameter from sao_band_filter

Pass instead the two variables from the struct needed in the function.
This simplifies writing asm optimized versions of the function

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

hevcdsp: remove unused parameter from sao_band_filter

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

avformat/utils: Fix number suffixes in tb_unreliable()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/h261enc: Fix number suffix

Found-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/h261dec: Fix exported MVs for skipped MBs

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/nellymoserenc: replace assert by av_assert1()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avfilter/vf_dctdnoiz: Use av_malloc_array()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Support YA16 in jpeg 2000.

avformat/oggenc: Simplify by using OFFSET and PARAM

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '60392480181f24ebf3ab48d8ac3614705de90152'

* commit '60392480181f24ebf3ab48d8ac3614705de90152':
  mem: fix pointer pointer aliasing violations

Conflicts:
libavutil/mem.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'f726fc21ef76a8ba3445448066f7b2a687fbca16'

* commit 'f726fc21ef76a8ba3445448066f7b2a687fbca16':
  ogg: Provide an option to offset the serial number

Conflicts:
libavformat/oggenc.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/motion_est: remove unneeded #include, there are no assert() only av_assert*

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/put_bits:  remove unneeded #include, there are no assert()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

mem: fix pointer pointer aliasing violations

This uses explicit memory copying to read and write pointer to pointers
of arbitrary object types. This works provided that the architecture
uses the same representation for all pointer types (the previous code
made that assumption already anyway).

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

ogg: Provide an option to offset the serial number

The ogg serial number doubles as codec id and sequence
value for concatenated samples.

dvenc: fix 2x4x8 (interlaced) weight table

Since the decoder has been fixed to output something similar to
other implementations, the encoder weight table needed this fix.

Reference PSNR values on a progressive sequence (from tiny_psnr)
from a chained encoding and decoding:
Full progressive: stddev:    0.74 PSNR: 50.69 MAXDIFF:   19
Before:           stddev:    1.43 PSNR: 44.97 MAXDIFF:   17
This patch:       stddev:    0.76 PSNR: 50.44 MAXDIFF:   14

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

examples/demuxing_decoding: set stream_idx in open_codec_context only if no error occured

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/motion_est: Set subcmp consistently for H261

Fixes Ticket3150

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/h261enc: More specific return code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/h261enc: Avoid casts, Simplify code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/cache: pass options to the underlying protocol via the url_open2

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge remote-tracking branch 'cehoyos/master'

* cehoyos/master:
  lavc/arm: Use the neon vertical chroma loop filter also for H.264 4:2:2.
  lavc/aarch64: Do not use the neon horizontal chroma loop filter for H.264 4:2:2.

Merged-by: Michael Niedermayer <michaelni@gmx.at>

configure: remove mp_filter leftover

Pointed out by: ubitux
Signed-off-by: Paul B Mahol <onemda@gmail.com>

avfilter/avfilter: Remove CONFIG_MP_FILTER case

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avfilter: remove vf_mp.c

Signed-off-by: Paul B Mahol <onemda@gmail.com>

remove libmpcodecs

Signed-off-by: Paul B Mahol <onemda@gmail.com>

doc/indevs/libcdio: apply minor spell fixes, extend documentation

lavd/libcdio: add more paranoia mode constants

lavd/libcdio: apply minor fixes to options documentation

lavc/arm: Use the neon vertical chroma loop filter also for H.264 4:2:2.

lavc/aarch64: Do not use the neon horizontal chroma loop filter for H.264 4:2:2.

avformat/movenc: remove unneeded #include, there are no assert() only av_assert*

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/mov: remove unneeded #include, there are no assert()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/mux: remove unneeded #include, there are no assert() only av_assert*

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/asfenc: remove unneeded #include, there are no assert() only av_assert*

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/mpeg: remove unneeded #include, there are no assert() only av_assert*

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/utils: remove unneeded #include, there are no assert() only av_assert*

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/vc1: remove unneeded #includes, there are no assert() only av_assert*

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/svq1dec: remove unneeded #include, there are no assert()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/h261enc: fix dquant encoding

Fixes Ticket4111

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavcodec/h264: replace assert() by av_assert0()

also remove the now unneeded #include <assert.h>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/ratecontrol: replace asserts by av_asserts

Also remove the now unneeded #include <assert.h>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/vp3: use init_get_bits8()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Merge remote-tracking branch 'cehoyos/master'

* cehoyos/master:
  Respect horizontal differencing predictor for 16bit gray tiff images.

Merged-by: Michael Niedermayer <michaelni@gmx.at>

lavfi: remove mp=softpulldown

It is supposed to work correctly with MEncoder only.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/ppc/h264qpel_template: Change if DEBUG assert() to av_assert2()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ppc/mpegvideo_altivec: Change assert to av_assert2()

And move it to where the variable is used

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/videodsp_template: remove unneeded #include, there are no assert() only av_assert*

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/qdm2: remove unneeded #include, there are no assert()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

ffmpeg: remove unused header

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/rtpdec_qdm2: change assert to av_assert0()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

ffmpeg: free_input_threads requires HAVE_PTHREADS

Hide behind that macro to avoid warning on systems with
!HAVE_PTHREADS.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Respect horizontal differencing predictor for 16bit gray tiff images.

Fixes ticket #4291.

doc/muxers/segment: document strftime option

It was added in commit 202947a0665ea523022afb0a6c50eed96bcd6b69.

avformat/nut: support WavPack

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/nut: add GIF[0]

Signed-off-by: Paul B Mahol <onemda@gmail.com>