For 422 frames we should not use hard coded 8 to calculate mb size for
uv plane. Chroma shift should be taken into consideration to be
compatiple with different sampling format.
The error is reported by fate test when av_cpu_max_align() return 64
on the platform supporting AVX512. This is a hidden error and it is
exposed after commit 17a59a634c.
mpeg2enc has a mechanism to reuse frames. When it computes SSE (sum of
squared error) on current mb, reconstructed mb will be wrote to the
previous mb space, so that the memory can be saved. However if the align
is 64, the frame is shared in somewhere else, so the frame cannot be
reused and a new frame to store reconstrued data is created. Because the
height of mb is wrong when compute sse on 422 frame, starting from the
second line of macro block, changed data is read when frame is reused
(we need to read row 16 rather than row 8 if frame is 422), and unchanged
data is read when frame is not reused (a new frame is created so the
original frame will not be changed).
That is why commit 17a59a634c exposes this
issue, because it add av_cpu_max_align() and this function return 64 on
platform supporting AVX512 which lead to creating a frame in mpeg2enc,
and this lead to the different outputs.
Signed-off-by: Wenbin Chen <wenbin.chen@intel.com>
Signed-off-by: Marton Balint <cus@passwd.hu>
This is more spec-compliant because it does not rely
on dead-code elimination by the compiler. Especially
MSVC has problems with this, as can be seen in
https://ffmpeg.org/pipermail/ffmpeg-devel/2022-May/296373.html
or
https://ffmpeg.org/pipermail/ffmpeg-devel/2022-May/297022.html
This commit does not eliminate every instance where we rely
on dead code elimination: It only tackles branching to
the initialization of arch-specific dsp code, not e.g. all
uses of CONFIG_ and HAVE_ checks. But maybe it is already
enough to compile FFmpeg with MSVC with whole-programm-optimizations
enabled (if one does not disable too many components).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This commit removes the ineffective FF_MPV_DEPRECATED_ options,
namely mpeg_quant (this is only an option for MPEG-4), a53cc
(this is only an option for MPEG-2), force_duplicated_matrix
(applies only to MJPEG) and b_strategy, b_sensitivity and brd_scale
(these options only make sense for encoders supporting B-frames,
which currently means the MPEG-1/2 and MPEG-4 encoders).
Given that these options never changed the outcome of encoding,
they are removed at once.
Notice that the options for the encoders for which it made sense
are not affected by this.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
(M)JPEG does not use motion estimation/motion vectors at all.
These options therefore don't affect the output at all.
So remove them.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
It is a remnant of the old way for user-supplied buffers;
it is always-true since 93016f5d1d.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
If one encodes MJPEG with a single slice and uses input with
AV_FRAME_DATA_ICC_PROFILE side data, the current allocation code
in ff_mpv_encode_picture() will always increase the size of the
temporary buffer used for allocating packets by the size needed
for to write the ICC chunk even when the current buffer is actually
large enough. This commit fixes this.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
MJPEG is the only mpegvideo-based encoder making use of it.
Fixes linking failures in case mpegvideo_enc.c is compiled
with AMV, LJPEG and MJPEG encoders disabled.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This is mostly straightforward. The major complication is that, as a
result of the 16-bit chunk size limitation, ICC profiles may need to be
split up into multiple chunks.
We also need to make sure to allocate enough extra space in the packet
to fit the ICC profile, so modify both mpegvideo_enc.c and ljpegenc.c to
take into account this extra overhead, failing cleanly if necessary.
Also add a FATE transcode test to ensure that the ICC profile gets
written (and read) correctly. Note that this ICC profile is smaller than
64 kB, so this doesn't test the APP2 chunk re-arranging code at all.
Signed-off-by: Niklas Haas <git@haasn.dev>
It is currently a "Picture", an mpegvideo-specific type
that has a lot of baggage, all of which is unnecessary
for new_picture, because only its embedded AVFrame
is ever used. So just use an ordinary AVFrame.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
One can use slices without slice-threading. The results for
mpegvideo-encoders are abysmal: AMV, SpeedHQ, H.263, RV10, RV20,
MSMPEG4v2, MSMPEG4v3 and WMV1 produce broken files.
WMV2 meanwhile expects the MpegEncContext given to ff_wmv2_encode_mb()
to be at the beginning of a Wmv2Context (a structure that this encoder
shares with the WMV2 decoder), yet this is only true for the
main context and not for the slice contexts, leading to segfaults.
SpeedHQ additionally triggers an av_assert2, because it is not
byte-aligned at a position where it ought to be byte-aligned.
Given that no codec not supporting slice threading works this commit
disallows using slices unless the encoder supports slice threading.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This avoids unnecessary rebuilds of most source files if only the
list of enabled components has changed, but not the other properties
of the build, set in config.h.
Signed-off-by: Martin Storsjö <martin@martin.st>
Seems to have been always unused since these functions were introduced
in 1f0cd30fd9.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
mpegvideo-based encoders supporting bframes implement this
by opening encoders of their own to test how long the chains
of bframes are supposed to be. The needed AVCodec was obtained
via avcodec_find_encoder(). This is complicated, as the current
encoder can be directly obtained. And it also is not guaranteed
that one actually gets the current encoder or not another encoder
for the same codec ID (the latter does not seem to be the case now).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
An offset has the advantage of not needing to be updated
when the buffer is reallocated. Furthermore, the way the pointer
is currently updated is undefined behaviour in case the pointer
is not already set (i.e. when not encoding MPEG-1/2), because
it calculates the nonsense NULL - s->pb.buf.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Also use said function in mpegvideo.c and mpegvideo_enc.c;
and make ff_free_picture_tables() static as it isn't needed anymore
outside of mpegpicture.c.
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The current size is AV_NUM_DATA_POINTERS (i.e. eight).
This number is chosen in order to minimize the amount of allocations
for AVFrame.extended_(data|buf) for audio; it is meaningless
for video for which four is sufficient. So decrease this array
in order to minimize what is copied in ff_mpeg_ref_picture()
and at the places that copy a whole MpegEncContext.
Also do the same for snowenc.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
encode_mb() calls encode_mb_internal() three times, once
for each supported chroma format. The reason for this is
that some chroma format dependent parameters can then be
inlined as encode_mb_internal() is marked as av_always_inline.
Yet the most basic parameters based upon chroma format have
not been inlined: The chroma format itself and the chroma
subsampling parameters. This commit does so.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This factors the translation from MpegEncContext out
and will enable further optimizations in the next commits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The generic code ensures that only codecs with
the FF_CODEC_CAP_AUTO_THREADS internal cap ever have to
handle the case avctx->thread_count == 0 themselves;
moreover, it is also ensured generically that only codecs
that support some form of threading have thread_count set
to something else than one. So these checks are unnecessary.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Currently, ff_mpv_encode_end() is the close function of
the two MJPEG-based encoders; it calls ff_mjpeg_encode_close()
for them which adds a check to the generic code.
This commit reverses the order of this relationship:
The MJPEG encoders directly use a custom close function
which in turn calls ff_mpv_encode_end(). This avoids the branch
in ff_mpv_encode_end() and makes the generic code more generic.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Wenbin Chen
Andreas Rheinhardt
Niklas Haas
Martin Storsjö