avcodec/jpeg2000dwt: Replace /2 by >>1

Divisions can be slow if the compiler fails to replace them by shifts

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

swscale/x86/rgb2rgb_template: fix signedness of v in shuffle_bytes_2103_{mmx,mmxext}

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>

swscale/x86/rgb2rgb_template: don't call emms on sse2/avx functions

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>

swscale/x86/rgb2rgb_template: add missing xmm clobbers

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>

avcodec/texturedspenc: Add () to protect macro / argument evaluation order

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

vaapi_h264: fix RefPicList[] field flags.

Use new H264Ref.reference field to track field picture flags. The
H264Picture.reference flag in DPB is now irrelevant here.

This is a regression from git commit d8151a7, and that affected
multiple interlaced video streams.

Signed-off-by: Gwenole Beauchesne <gwenole.beauchesne@intel.com>

vda: unlock the pixel buffer base address.

The pixel buffer base address is never unlocked this causes
a bug with some pixel format types that are produced natively
by the hardware decoder: the first buffer was always used.
Unlock the pixel buffer base address fixes the issue.

movtextdec.c: Improve upon dynarrays and text_to_ass

Signed-off-by: Niklesh <niklesh.lalwani@iitb.ac.in>

swscale/rgb2rgb_template: Fix signedness of v in shuffle_bytes_2103_c()

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

swscale/rgb2rgb_template: Implement shuffle_bytes_0321_c and fix shuffle_bytes_2103_c on BE

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

avcodec/texturedsp: Add protective () to RGBA() macro

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

avformat/mpegts: Use STREAM_TYPE_PRIVATE_DATA instead of 6

Suggested-by: Wolfgang Lorenz <wl-chmw@gmx.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/electronicarts: Fixed ea_probe function to accept vp6a videos

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

tests/fate/image: also run fate-sgi for the fate-image target

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

Changelog: Move HAP to the correct section

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

avcodec/dds: Fix palettes on big endian

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

Merge commit '5c018ee18895f88e9e1d2174059dcdd48bf872d2'

* commit '5c018ee18895f88e9e1d2174059dcdd48bf872d2':
  DirectDraw Surface image decoder

Conflicts:
Changelog
configure
doc/general.texi
libavcodec/Makefile
libavcodec/version.h
tests/fate/image.mak

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

avcodec/jpeg2000dwt: Move large arrays used in the test code away from the stack

This should avoid problems on systems with little stack space and fix some crashes
in fate

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

avcodec/jpeg2000dwt: Print 1 digit less in the 9/7f DWT test

This avoids test failure due to differing rounding between 32 and 64bit x86

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

avcodec/jpeg2000dwt: Use a tighter check threshold for the 9/7f DWT test

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

Merge commit '7ca3e5203f133eb41a0b5c3a1d753a7427ba72e7'

* commit '7ca3e5203f133eb41a0b5c3a1d753a7427ba72e7':
  Hap decoder and encoder

Conflicts:
Changelog
configure
doc/general.texi
libavcodec/allcodecs.c
libavcodec/version.h

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

swscale/rgb2rgb_template: Disable shuffle_bytes_2103_c on big endian

The function is specific to little endian

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

Merge commit 'c0b105756f61d253bdabcc2bb49453a2557e7c3b'

* commit 'c0b105756f61d253bdabcc2bb49453a2557e7c3b':
  txd: Use the TextureDSP module for decoding

Conflicts:
configure
libavcodec/s3tc.c
libavcodec/s3tc.h
libavcodec/txd.c

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

Merge commit '8337e0c57345f24cf6471220e5f8a0ea21b7c1d0'

* commit '8337e0c57345f24cf6471220e5f8a0ea21b7c1d0':
  Introduce a TextureDSP module

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

Merge commit '2ecfd451649c7a08cb633635df98e59f7c6e2140'

* commit '2ecfd451649c7a08cb633635df98e59f7c6e2140':
  Implement Snappy decompression

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

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for block functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for block functions in new file blockdsp_msa.c

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

fate: add some tests for ffv1 level 3 with 8/10/16 bps

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

swr: Remember previously set int_sample_format from user

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

swresample/swresample: Clear delayed_samples_fixup in clear_context()

This probably makes no difference but its more proper

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

DirectDraw Surface image decoder

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

Hap decoder and encoder

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

txd: Use the TextureDSP module for decoding

Using the internal DXTC routines brings support for non multiple of 4
textures. A new test is added to cover this feature. Hashes differ
since the decoding algorithm is different, though no visual changes
have been spotted.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

Introduce a TextureDSP module

This module implements generic texture decompression from different
families (DXTC, RGTC, BCn) and texture compression DXTC 1, 3, and 5.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

Implement Snappy decompression

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>

lavf/file: implement move and delete callbacks

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

lavf/avio: Extend API with avio_move() and avio_delete()

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

Merge commit '61dc9d647c6664e11674d9a10fdde29987d6acda'

* commit '61dc9d647c6664e11674d9a10fdde29987d6acda':
  udp: Fix local_port management

Conflicts:
libavformat/udp.c

See: 4f3f5ee1ac93797954eb66ef4d19e2256c0aaee7
Merged-by: Michael Niedermayer <michaelni@gmx.at>

udp: Fix local_port management

The default value for unset is -1, not 0.

Problem introduced in 66028b7ba6b411ba12ef553e9c8f1f4a4fe27710

avcodec/jpeg2000: Move L band scaling from the 9/7f wavelet to quantization stage

This reduces the number of operations
Its not done for 9/7i as that would overflow thanks to JPEG2000 allowing
32 decomposition levels

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

avcodec/jpeg2000dwt: use 32x32->64 multiplies in the 9/7i DWT

This significantly improves the quality when the integer 9/7 transform
is used

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

avcodec/jpeg2000: Move H band scaling from wavelet into quantization code

This reduces the number of operations done and is equivalent except for
rounding

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

avcodec/jpeg2000dwt: Move K/X constants to header

this makes them accessible by the rest of the jpeg2000 code

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

avcodec/jpeg2000: Move gainb handling into the quantization code

thats how the specification defines it, this also improves numerical
accuracy of the integer wavelet implementation. It otherwise should
be equivalent, in case of overflows this can be reverted.

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

ffserver: drop unneeded else branching

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffserver: drop unneeded else branching

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffserver: drop unneeded else branching

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

ffserver: formating

Signed-off-by: Reynaldo H. Verdejo Pinochet <reynaldo@osg.samsung.com>

swresample: soxr implementation for swr_get_out_samples()

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

Merge commit '8655c54160767de1a6b96f8bc310d6e4eaceff48'

* commit '8655c54160767de1a6b96f8bc310d6e4eaceff48':
  libvpx: Support the vp9 extended profiles

Conflicts:
Changelog
configure
libavcodec/libvpx.c
libavcodec/libvpx.h
libavcodec/libvpxdec.c
libavcodec/libvpxenc.c
libavcodec/version.h

mostly not merged, as requested by jamrial

See: 01e59d48ed1a41b88107ed1d4d56ae0cbcd1a60e
See: 079b7f6eacc09bc2813fc1ddc230ab05022b69c2 and others
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '05bf3f54e6078716f6267df530bf1d474ca3d606'

* commit '05bf3f54e6078716f6267df530bf1d474ca3d606':
  libvpx: Do not set vp8 only parameters when encoding in vp9

Conflicts:
libavcodec/libvpxenc.c

See: 238ec505e263e7d0363798a05c1c28f494a6300a
See: ae330070ee8b985abb011835502b03baacfbb18c
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'c060d046aa2f89c0e601a2dcfbce53f0e36cf498'

* commit 'c060d046aa2f89c0e601a2dcfbce53f0e36cf498':
  af_resample: Set the number of samples in the last frame

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

Merge commit '6ec688e1bc76dd93151cbca1c340162ae4b10d77'

* commit '6ec688e1bc76dd93151cbca1c340162ae4b10d77':
  mp3: enable packed main_data decoding in MP4

Conflicts:
libavcodec/mpegaudiodec_template.c

Only the parts needed to support the available sample are merged
the remaining error checks are left in place

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

avcodec/jpeg2000dwt: Fix value of F_LFTG_X and I_LFTG_X

Even if the jpeg2000 spec uses a wrong value this does not
make mathematics work this way, also this has been corrected in the 2004
version AFAIK

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

swresample/swresample: Print used int_sample_fmt

Suggested-by: wm4
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

swresample: Choose 16bit internally only if input and output is 16bit or less

or if no rematrix and no resampling is performed and the input is 16bit
note reampling and rematrix itself always use more than 16bit internally
the "internal" sampling format is the format between these steps

Its unlikely the difference from this commit is audible in any case
unless there is some bug either before or after the change.
but multiple people prefer this and it slightly improves the precission
of computations.

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

avcodec/jpeg2000dwt: also test 9/7 float DWT

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

libvpx: Support the vp9 extended profiles

Bump the minimum libvpx version to 1.3.0 and rework the configure logic
to fail only if no decoders and encoders are found.

Based on the original patch from Vittorio.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

avcodec/jpeg2000dwt: List differences for DWT test

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

avcodec/jpeg2000dwt: increase the number of test iterations

the code had too little coverage

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

avcodec/jpeg2000dwt: Also test 9/7 integer DWT

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

avcodec/jpeg2000dwt: Allow testing with a maximum allowed difference

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

avformat/brstm: support little-endian s16 PCM

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

lavc/adpcm: THP: don't use the ADPC/SEEK table when not seeking

This is almost certainly closer to how the actual Nintendo players work,
and fixes some output pops in files with blank ADPC/SEEK tables (like
those from brawlcustommusic).

lavf/brstm: handle a BFSTM endianness oddity

lavf/brstm: add support for seeking

lavf/brstm: cleanup; fix short-block demuxing

lavf/brstm: move bfstm var to a local

avformat/rawenc: Store sample number for ADX

Fixes Ticket4540

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

Support demuxing 4gv codec stored in qcp files

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

ffmpeg_opt: Fix handling of creation_time now

Fixes Ticket4495

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

avformat/movenc: Check return code of ff_iso8601_to_unix_time()

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

avcodec/mjpegdec: dont try to combine fields for decimated multiscope 2 material

Fixes Ticket4535

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

libvpxenc: support setting colorspace for vp9

the vp9 bitstream supports 8 values:
unknown (default), bt601, bt709, smpte170, smpte240, bt2020, reserved
and sRGB.

Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: James Zern <jzern@google.com>

lavc/adpcm: THP: fix indentation

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

avcodec/mqcenc: Remove unused variable

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

libavfilter/formats: Fix parsing of channel specifications with a trailing 'c'.

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

lavc/adpcm: THP: allow channel counts up to 10

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

lavf/brstm: allow larger block sizes

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

lavf/brstm: support little-endian files (BCSTM; 3DS)

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

lavc/adpcm: THP: handle packets with sample counts not divisible by 14

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

lavc: add little-endian ADPCM_THP decoder

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

fate: add avfilter/formats-test

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

avfilter/formats: Add test for ff_parse_channel_layout()

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

avcodec/jpeg2000dec: Replace TPsot check by assert

The condition is impossible now.
the TPsot type is changed to unsigned int to prevent gcc from producing warnings

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

avcodec/jpeg2000dec: Increase tile part limit to 256

This is the maximum that the syntax of jpeg2000 is capable to store

Fixes Ticket4639

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

libvpx: Do not set vp8 only parameters when encoding in vp9

af_resample: Set the number of samples in the last frame

Otherwise trailing zeroes would appear.

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

mp3: enable packed main_data decoding in MP4

14496-3 suggests packing main_data of MP3 that is usually scattered
into multiple frames due to bit reservoir.

However, after packing main_data into a access unit, bitrate index
in the MPEG audio frame header doesn't match with actual frame size.

In order to accept this, this patch removes unnecessary frame size
checking on mp3 decoder.

Also, mov demuxer was changed to use MP3 parser only on special cases
(QT MOV with specific sample description) to avoid re-packetizing.

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

avutil/fixed_dsp: remove ff_ prefix from static function

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

avcodec/jpeg2000dwt: Fill array using 2d instead of 1d accesses

Avoids gcc warning

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

avcodec/jpeg2000dwt: Factor dwt test code out so it can be easily used with the other dwts too

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

avcodec/jpeg2000dwt: Fix  special cases in sd_1d53()

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

tests/fate: Add fate-j2k-dwt

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

avcodec/jpeg2000dwt: add test for the 5/3 wavelet

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

avcodec/jpeg2000dwt: Fix order of operations in integer decomposition wavelets

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

avcodec/j2kenc: Use JP2 format by default

Without the JP2 fields its not possible to identify the colorspace

Fixes Ticket538

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

avcodec/j2kenc: Support writing JP2 encapsulation

The added option matches openjpeg

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

avcodec/dpxenc: implement write16/32 as functions

Fixes undefined behavior and segfault

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

avutil/avstring: Do not print NULL

Fixes segfault
Fixes Ticket4452

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

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for hpel functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for hpel functions in new file hpeldsp_msa.c
Adds new generic macros (needed for this patch) in libavutil/mips/generic_macros_msa.h

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>