Improves compatibility with Resolume products by adding an additional
hashtable for DXT color+LUT combinations, and padding the DXT texture
dimensions to the next largest multiple of 16. Produces identical
packets to Resolume Alley in manual tests.
Signed-off-by: Emma Worley <emma@emma.gg>
The follow cmd output corrupted file before the patch:
ffmpeg -f lavfi -i color=blue,trim=duration=0.04 \
-f lavfi -i anullsrc,atrim=duration=2 \
-movflags +empty_moov+hybrid_fragmented \
-frag_duration 1000000 \
-frag_interleave 1 \
output.mp4
1. first_track is the first track with track->entry != 0. As in the
command above, video track (track index 0) has a single frame. When
flush the second fragment, first_track is 1, the audio track.
2. write_moof = i == first_track, so write_moof is false for i = 0.
3. When mov->frag_interleave != 0, mov->mdat_buf != NULL, because
it contains audio data. So avio_write is called before write_moof,
that is, the data write before moof, and mov_finish_fragment
executed with wrong mdat_start.
4. With normal fmp4 output, the error isn't obvious. With
hybrid_fragmented, ffplay output.mp4 shows a lot of error messages.
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
Up until now, our muxer wrote FFV1 in video-for-windows
compatibility mode out of concern for old demuxers that
only support that (whereas the demuxer accepts V_FFV1).
This commit switches to using native mode, because
a) V_FFV1 is around long enough so that old demuxers
should not be an issue (support in FFmpeg has been added
in commit 9ae762da7e
in March 2017/FFmpeg 3.3),
b) using native mode uses fewer bytes for the CodecPrivate,
c) the VfW extradata is zero-padded to an even length
if necessary, but our demuxer forgot to undo the padding
until very recently (92e310eb82),
so that there are many versions of our demuxer around that
are buggy wrt VFW, but not V_FFV1.
This affects the FFV1 extradata checksums, specifically
the (experimental) version 4 files with error check version 2*
as created by
ffmpeg -i ../fate-suite/mpeg2/sony-ct3.bs -c:v ffv1 \
-slices 16 -frames 1 -level 4 -strict experimental ffv1.mkv
VFW files like the above created by this muxer before this patch
would not work with an old demuxer.
*: Without error check version 2, the CRC for the whole extradata
is zero, which is not changed by appending a zero byte.
Reviewed-by: compn <ff@hawaiiantel.net>
Reviewed-by: Dave Rice <dave@dericed.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The structure is padded to an even length with an internal
size field to indicate the real size.
The matroska-matroska-display-metadata test (writing FFV1
in VFW mode) was affected by this.
It should also fix ticket #11613.
Reviewed-by: compn <ff@hawaiiantel.net>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Up until now, the encoder replicated all the border pixels
for incomplete 16x16 macroblocks. In case the available width
or height is <= 8, some of the luma blocks of the MB
do not correspond to actual input, so that we should encode
them using the least amount of bits. Zeroing the block coefficients
(as this commit does) achieves this, replicating the pixels
and performing an FDCT does not.
This commit also removes the frame copying code for insufficiently
aligned dimensions.
The vsynth3-asv[12] FATE tests use a 34x34 input file and are
therefore affected by this. As the ref updates show, the size
and checksum of the encoded changes, yet the decoded output
stays the same.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Don't use a 7.1 EAC3 input file for which our decoder is not
bitexact; instead just use the asynth-44100-8.wav file
which (as a 7.1 file) exhibits the same issue fixed by
1b3f4842c1.
(Either the encoder or the resampler are still not completely
bitexact, so we limit the number of frames output.)
Also switch to a framecrc test so that the output channel layout
is directly contained in the ref file.
Reviewed-by: James Almer <jamrial@gmail.com>
Reviewed-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The file has buggy timestamps (it uses B-frames, yet pts==dts)
and therefore the last frame is currently discarded by FFmpeg cli.
Using -fps_mode passthrough avoids this and provides coverage
of the SVQ3 draining logic.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
I have several .ts captures where video and audio codec changes even
though the PMT version does not change and the PIDs stay the same.
This happens during transition to/from slate (mpeg2 video and audio)
to network broadcast (hevc video and eac3 audio in private PES).
I've updated fate ts-demux expected results.
Bitstream generated using the reference encoder, then edited to fix the
colour description and an extra metadata block added. FFmpeg decoder
output is identical to the reference decoder output.
The content used is the first three frames of "Waterfall" from the SVT
Open Content Video Test Suite 2022. This is copyright Sveriges
Television AB and is used under the Creative Commons Attribution 4.0
International License.
This bsf converts AV_PKT_DATA_NEW_EXTRADATA side data in avcc format
to in-band annexb format. However, the side data wasn't been removed
and copied from input packet to output packet. So the output packet
has mixed bitstream format. We don't support mixed bitstream format.
For example, h264_metadata report error in the following case:
ffmpeg -i foo.flv \
-bsf:v "h264_mp4toannexb,h264_metadata" \
-c copy -f null
This patch removed NEW_EXTRADATA side data after process.
This patch also add a check so only NEW_EXTRADATA in avcc format is
processed. NEW_EXTRADATA in annexb format is copied to output as is.
Reported-by: jiangjie <jiangjie618@gmail.com>
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
The earlier code only used the counts from the last slice.
The two FATE tests using slices show compression improvements
due to this.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
It is unnecessary, as we already have the entries sorted by
probability and therefore implicitly by length. All we need
on top of that to build the tree is the number of entries
of a given length.
Doing so gives a 3.6% speedup of ff_mjpeg_encode_huffman_close()
here; it also saves about 640B of .text here.
The new code puts values with higher probability to the left
of the tree. The old code did not and therefore
the FATE checksums needed to be updated. Due to MJPEG's
0xFF unescaping file sizes as well as file checksums
needed to be updated; the decoded picture hashes stayed
the same. Given that codes on the left of the tree have
on average fewer bits set than codes on the right, the
file sizes mostly improve (all except vsynth3-mjpeg-444).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
A sample with a particular partitioning structure that could not be read
correctly before 26c5d8cf5d
Signed-off-by: Frank Plowman <post@frankplowman.com>
Signed-off-by: James Almer <jamrial@gmail.com>
