This fixes mixing up contexts, use of uninitialized data and crashes.
More specifically:
==1001752== Conditional jump or move depends on uninitialised value(s)
==1001752== at 0xA9ED82: avpriv_h264_has_num_reorder_frames (h264dec.c:64)
==1001752== by 0x668C7E: has_decode_delay_been_guessed (demux.c:757)
==1001752== by 0x66AB13: compute_pkt_fields (demux.c:1137)
==1001752== by 0x66B2E9: parse_packet (demux.c:1265)
==1001752== by 0x66BD84: read_frame_internal (demux.c:1449)
==1001752== by 0x67085B: avformat_find_stream_info (demux.c:2692)
==1001752== by 0x25157C: ifile_open (ffmpeg_demux.c:1814)
==1001752== by 0x272B15: open_files (ffmpeg_opt.c:1366)
==1001752== by 0x272D85: ffmpeg_parse_options (ffmpeg_opt.c:1415)
==1001752== by 0x2925C9: main (ffmpeg.c:991)
==1001752== Uninitialised value was created by a heap allocation
==1001752== at 0x483E0F0: memalign (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==1001752== by 0x483E212: posix_memalign (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==1001752== by 0x14882CE: av_malloc (mem.c:107)
==1001752== by 0x1463785: av_buffer_alloc (buffer.c:82)
==1001752== by 0x146423F: pool_alloc_buffer (buffer.c:369)
==1001752== by 0x14643C4: av_buffer_pool_get (buffer.c:407)
==1001752== by 0x752C4B: buffer_pool_get (mpegts.c:1142)
==1001752== by 0x7538F2: mpegts_push_data (mpegts.c:1407)
==1001752== by 0x758893: handle_packet (mpegts.c:2909)
==1001752== by 0x758E90: handle_packets (mpegts.c:3048)
==1001752== by 0x759B1D: mpegts_read_packet (mpegts.c:3290)
==1001752== by 0x6687A3: ff_read_packet (demux.c:649)
==1001752== by 0x66B594: read_frame_internal (demux.c:1346)
==1001752== by 0x67085B: avformat_find_stream_info (demux.c:2692)
==1001752== by 0x25157C: ifile_open (ffmpeg_demux.c:1814)
==1001752== by 0x272B15: open_files (ffmpeg_opt.c:1366)
==1001752== by 0x272D85: ffmpeg_parse_options (ffmpeg_opt.c:1415)
==1001752== by 0x2925C9: main (ffmpeg.c:991)
Found-by: Alexander A. Shvedov <shvedov@gmx.com>
CC: Pavel Koshevoy <pkoshevoy@gmail.com>
This reverts commit 0021484d05.
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
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.
Only those that can be decoded with our implementation, so excluding
* C and D - independent layers
* G, H, I - more than 2 layers
Frame hashes verified against the reference implementation from
https://hevc.hhi.fraunhofer.de/svn/svn_3DVCSoftware/
The test constructs 66 sine waves of different frequencies and joins
them into a 7th order ambisonic layout with extra non-diegetic
stereo. The channels are then shifted circularly with channelmap and
separated into individual streams for muxing with channelsplit.
Note that the last step is necessary because swr is not capable of
interleaving planar to packed for more than 64 channels, which would be
necessary to mux such a stream as PCM.
Fix "ost->st->avg_frame_rate = ost->frame_rate" in streamcopy_init()
being reset to input's frame rate a few lines below.
Note that in current code, there are some discrepancies amongst the
muxers. For example, avienc relies on time_base, so it is not affected
by this patch, whereas mxfenc and matroskaenc do use avg_frame_rate,
so this patch fixes -r being honored.
In the updated fate test, the input is (wrongly) probed as 50fps. With
this patch, the correct value (25fps) is successfully forced with -r.
Signed-off-by: Nicolas Gaullier <nicolas.gaullier@cji.paris>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
The time_base was a bad guess.
Currently, fate-time_base test data assumed that overriding the input
time_base would affect the frame_rate, but this behaviour is not
documented, so just fix the fate data now that this is fixed.
Fix regression since 10185e2d4c:
previously, when streamcopying, the time_base was guessed from the
frame_rate considering it is often constant, so guessing the frame_rate
back from the time_base was often not a problem.
To reproduce:
ffmpeg -i fate-suite/mpeg2/dvd_still_frame.vob -an -c copy out.mxf
Signed-off-by: Nicolas Gaullier <nicolas.gaullier@cji.paris>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
Replace the manually specified chroma location by one using standard
notation, arbitrarily "bottomleft" as it is a less common path.
Required if we want to phase out the use of manual chroma locations.
Encoders may emit a buffering period SEI without a corresponding
SPS/PPS if the SPS/PPS is carried out-of-band, eg with avcc.
During Annex B conversion, this may result in the SPS/PPS being
inserted *after* the buffering period SEI but before the IDR NAL.
Since the buffering period SEI references the SPS, the SPS/PPS
needs to come first.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
This test confirms that we can write mDCv and cLLi chunks and read them
back via the png decoder. It uses an HEVC conformance sample with this
metadata as the base source for the side data in the frames.
Signed-off-by: Leo Izen <leo.izen@gmail.com>
Reported-by: Jan Ekström <jeebjp@gmail.com>
Reviewed-by: Jan Ekström <jeebjp@gmail.com>
Reviewed-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This lets us detect when a container has flagged a stream as multilayer.
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
Do not pass an options dictionary to the avcodec_open2() in enc_open().
This is cleaner and more robust, as previously various bits of code
would try to interpret the contents of the options dictionary, with
varying degrees of correctness. Now they can just access the encoder
AVCodecContext directly.
Cf. 372c78dd42 - analogous change for
decoding.
A non-progressive field order is now written on the container level in
interlaced ProRes encoding tests.
Perform av_clip_int16(val) _after_ copying the value to last_dc.
This change ensures that clipping is applied only within the context of
the current block, preventing the propagation of clipped values to
subsequent DC components.
Related commits: c28f648b19 and dffae122d0
Related ticket: 4683
This avoids hardcoding any implementation-specific limitiations as
part of the API, and allows for future expandability.
This also allows API users to more conveniently convert the
values into floats without hardcoding specific conversion constants.
The API was committed a few days ago, so changing this field now
is within the realms of acceptable.
This allows ending up with a normal, non-fragmented file when
the file is finished, while keeping the file readable if writing
is aborted abruptly at any point. (Normally when writing a
mov/mp4 file, the unfinished file is completely useless unless it
is finished properly.)
This results in a file where the mdat atom contains (and hides)
all the moof atoms that were part of the fragmented file structure
initially.
Signed-off-by: Martin Storsjö <martin@martin.st>
ffprobe is meant to generate parseable output, and if a field is present, it
should be printed even if it has a default value.
Signed-off-by: James Almer <jamrial@gmail.com>
These tables are supposed to contain the number of bits needed
to encode a given (run, level) pair. Yet the number of bits
for pairs needing the escape code was wrong (it only contained
the escape code and not the bits needed for run and level).
Furthermore, H.261 (a format with explicit end-of-block codes)
does not work well together with the RLTable API from rl.c:
The EOB code is the first one in ff_h261_rl_tcoeff's VLC table
and has a run value of zero. Therefore the result of get_rl_index()
is off by one for run == 0 and level values with explicit
(run, level) pair.
Fixing this necessitated changing the ref files of the
vsynth*-h261-trellis tests. Both filesizes as well as PSNR
decreased. If one used a qscale value of 11 for this test,
one would have received files with about the same size as
before this patch (with qscale 12), but with better PSNR.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Before After
-------------------------------------------------
make fate-vvc CPU Time (No ASM) 131.52s 134.83s
libavcodec/vvc/* Line Coverage 95.3% 96.9%
inter_template.c Line Coverage 74.3% 88.2%
inter.c Line Coverage 85.3% 99.2%
Signed-off-by: Frank Plowman <post@frankplowman.com>
Also make the iso_channel_position table consistent with what the AAC decoder
uses in avcodec/aac/aacdec_usac.c.
Fate changes are caused by the change of how 7.1 layout is mapped, previously
it included Side Surround channels, now it includes the Surround channels.
Signed-off-by: Marton Balint <cus@passwd.hu>
