It is done later in ff_mpv_frame_start() (and nobody uses
current_picture_ptr between setting it in ff_mpv_frame_start()).
(The reason the vsynth*-h263-obmc ref files change is because
the call to ff_find_unused_picture() now happens after the older
pictures have been unreferenced in ff_mpv_frame_start(),
so that their slots in the picture array can be immediately
reused; the obmc code is somehow buggy and changes its output
depending on the earlier contents of the motion_val buffer.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Provide optimized implementation of pix_abs8 function for arm64.
Performance comparison tests are shown below.
- pix_abs_1_0_c: 101.2
- pix_abs_1_0_neon: 22.5
- sad_1_c: 101.2
- sad_1_neon: 22.5
Benchmarks and tests are run with checkasm tool on AWS Graviton 3.
Signed-off-by: Martin Storsjö <martin@martin.st>
Provide optimized implementation of sse8 function for arm64.
Performance comparison tests are shown below.
- sse_1_c: 130.7
- sse_1_neon: 29.7
Benchmarks and tests run with checkasm tool on AWS Graviton 3.
Signed-off-by: Hubert Mazur <hum@semihalf.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
Provide optimized implementation of pix_abs16_y2 function for arm64.
Performance comparison tests are shown below.
pix_abs_0_2_c: 317.2
pix_abs_0_2_neon: 37.5
Benchmarks and tests run with checkasm tool on AWS Graviton 3.
Signed-off-by: Hubert Mazur <hum@semihalf.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
Provide neon implementation for sse4 function.
Performance comparison tests are shown below.
- sse_2_c: 80.7
- sse_2_neon: 31.0
Benchmarks and tests are run with checkasm tool on AWS Graviton 3.
Signed-off-by: Hubert Mazur <hum@semihalf.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
Provide neon implementation for sse16 function.
Performance comparison tests are shown below.
- sse_0_c: 268.2
- sse_0_neon: 43.5
Benchmarks and tests run with checkasm tool on AWS Graviton 3.
Signed-off-by: Hubert Mazur <hum@semihalf.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
Since d69d12a5b9 these av_assert2()
(or more exactly, the ones in hadamard8_diff8x8_c() and
hadamard8_intra8x8_c()) are hit. So just remove all of these asserts.
(If the test were improved to know which functions expect h == 8
and which support any value, the asserts could be readded
at the appropriate places.)
Reviewed-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The height is hardcoded in some of the me_cmp functions, but not
in all of them. But in the case of all other functions, it's hardcoded
in the same place in SIMD functions as in the C reference functions,
while this one function differs from the behaviour of the C code.
(Before 542765ce3e, there were a
couple other sad8_*_mmx functions with similar hardcoded height.)
Signed-off-by: Martin Storsjö <martin@martin.st>
mpegvideo uses an array of Pictures and when it is done with using
them, it only unreferences them incompletely: Some buffers are kept
so that they can be reused lateron if the same slot in the Picture
array is reused, making this a sort of a bufferpool.
(Basically, a Picture is considered used if the AVFrame's buf is set.)
Yet given that other pieces of the decoder may have a reference to
these buffers, they need not be writable and are made writable using
av_buffer_make_writable() when preparing a new Picture. This involves
reading the buffer's data, although the old content of the buffer
need not be retained.
Worse, this read can be racy, because the buffer can be used by another
thread at the same time. This happens for Real Video 3 and 4.
This commit fixes this race by no longer copying the data;
instead the old buffer is replaced by a new, zero-allocated buffer.
(Here are the details of what happens with three or more decoding threads
when decoding rv30.rm from the FATE-suite as happens in the rv30 test:
The first decoding thread uses the first slot of its picture array
to store its current pic; update_thread_context copies this for the
second thread that decodes a P-frame. It uses the second slot in its
Picture array to store its P-frame. This arrangement is then copied
to the third decode thread, which decodes a B-frame. It uses the third
slot in its Picture array for its current frame.
update_thread_context copies this to the next thread. It unreferences
the third slot containing the other B-frame and then it reuses this
slot for its current frame. Because the pic array slots are only
incompletely unreferenced, the buffers of the previous B-frame are
still in there and they are not writable; in fact the previous
thread is concurrently writing to them, causing races when making
the buffer writable.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
At this point active_thread_type is set iff active_thread_type
is set to FF_THREAD_FRAME iff AVCodecInternal.frame_thread_encoder
is set.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
If qsv hwdevice is available, use the mfxLoader handle in qsv hwdevice
to create mfx session. Otherwise create mfx session with a new mfxLoader
handle.
This is in preparation for oneVPL support
The following Cflags has been added to libmfx.pc, so mfx/ prefix is no
longer needed when including mfx headers in FFmpeg.
Cflags: -I${includedir} -I${includedir}/mfx
Some old versions of libmfx have the following Cflags in libmfx.pc
Cflags: -I${includedir}
We may add -I${includedir}/mfx to CFLAGS when running 'configure
--enable-libmfx' for old versions of libmfx, if so, mfx headers without
mfx/ prefix can be included too.
If libmfx comes without pkg-config support, we may do a small change to
the settings of the environment(e.g. set -I/opt/intel/mediasdk/include/mfx
instead of -I/opt/intel/mediasdk/include to CFLAGS), then the build can
find the mfx headers without mfx/ prefix
After applying this change, we won't need to change #include for mfx
headers when mfx headers are installed under a new directory.
This is in preparation for oneVPL support (mfx headers in oneVPL are
installed under vpl directory)
It is the proper place to set it, directly besides mb_width and
mb_stride. The reason for doing it the way it is done now seems
to be that the code does not create more slice contexts than necessary
(i.e. not more than one per row), so that this number needs to be
known before setting the number of slices. But this can always be
arranged by just moving the code that sets the number of slices.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
These fields are only ever set by the encoder for the current picture
and for no other picture. So only one set of these values needs to
exist, so move them to MpegEncContext.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Of all the buffers that are made writable, three are always allocated
and the other four are allocated iff any one of them is allocated;
so one can replace the seven checks for existence with one.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Up until now, ff_wmv2_decode_secondary_picture_header() only
set the mb_type array for non I-pictures, so that the decoding
process uses the earlier values of this array; this affects
the output of the wmv8-x8intra FATE-test (which this patch
therefore updates). These earlier values were set when decoding
earlier frames or when the buffer was initially zero-allocated.
A consequence of this is that the output of this test would be
random if ff_find_unused_picture() would select the unused picture
to return at random. Furthermore decoding from a keyframe onwards
depends upon the earlier state of the decoder.
This patch therefore zeroes said array when decoding an I picture.
(It is not claimed that zero is the right value to fill the array with.
I just don't know.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
These tables are only used by encoders and only for the current picture;
ergo they need not be put into the picture at all, but rather into
the encoder's context. They also don't need to be refcounted,
because there is only one owner.
In contrast to this, the earlier code refcounts them which
incurs unnecessary overhead. These references are not unreferenced
in ff_mpeg_unref_picture() (they are kept in order to have something
like a buffer pool), so that several buffers are kept at the same
time, although only one is needed, thereby wasting memory.
The code also propagates references to other pictures not part of
the pictures array (namely the copy of the current/next/last picture
in the MpegEncContext which get references of their own). These
references are not unreferenced in ff_mpeg_unref_picture() (the
buffers are probably kept in order to have something like a pool),
yet if the current picture is a B-frame, it gets unreferenced
at the end of ff_mpv_encode_picture() and its slot in the picture
array will therefore be reused the next time; but the copy of the
current picture also still has its references and therefore
these buffers will be made duplicated in order to make them writable
in the next call to ff_mpv_encode_picture(). This is of course
unnecessary.
Finally, ff_find_unused_picture() is supposed to just return
any unused picture and the code is supposed to work with it;
yet for the vsynth*-mpeg4-adap tests the result depends upon
the content of these buffers; given that this patchset
changes the content of these buffers (the initial content is now
the state of these buffers after encoding the last frame;
before this patch the buffers used came from the last picture
that occupied the same slot in the picture array) their ref-files
needed to be changed. This points to a bug somewhere (if one removes
the initialization, one gets uninitialized reads in
adaptive_quantization in ratecontrol.c).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Sufficiently recent Intel hardware is able to do encoding of 8bit 4:4:4
content in HEVC and VP9. The main requirement here is that the frames
must be provided in the AYUV format.
Enabling support is done by adding the appropriate encoding profiles
and noting that AYUV is officially a four channel format with alpha so
we must state that we expect all four channels.
vaapi_decode_find_best_format currently does not set the
VA_SURFACE_ATTRIB_SETTABLE flag on the pixel format attribute that it
returns.
Without this flag, the attribute will be ignored by vaCreateSurfaces,
meaning that the driver's default logic for picking a pixel format will
kick in.
So far, this hasn't produced visible problems, but when trying to
decode 4:4:4 content, at least on Intel, the driver will pick the
444P planar format, even though the decoder can only return the AYUV
packed format.
The hwcontext_vaapi code that sets surface attributes when picking
formats does not have this bug.
Applications may use their own logic for finding the best format, and
so may not hit this bug. eg: mpv is unaffected.
It (unfortunately) involves an allocation and can therefore fail.
Reviewed-by: Timo Rothenpieler <timo@rothenpieler.org>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The functions to replace parameter sets are only called
after the respective parameter set has just been read or
has just been written; all of these functions check
that the id field is within the appropriate range.
So the checks in the replace-functions can be removed.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
It is no longer used.
Also rename ff_cbs_alloc_unit_content2 to ff_cbs_alloc_unit_content.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
cbs_jpeg was the last user of CBS that didn't use
CodedBitstreamUnitTypeDescriptors.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
