RV10 and RV20 are unsupported because creating the correct CodecPrivate
is unsupported (the demuxer uses a codecpriv_offset of 26, so one
would need to recreate the missing 26 bytes); COOK and SIPR are
unsupported, because Matroska uses a packetization mode that is
different from what FFmpeg uses in its packets (see
matroska_parse_rm_audio() in the demuxer).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
bda44f0f39 added code that
potentially added another BlockMore master and BlockAdditional
data as well as BlockAddID number, yet it bumped the number
of EBML elements by four instead of only three.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Matroska supports orthogonal transformations (both pure rotations
as well as reflections) via its 3D-projection elements, namely
ProjectionPoseYaw (for a horizontal reflection) as well as
ProjectionPoseRoll (for rotations). This commit adds support
for this.
Support for this in the demuxer has been added in
937bb6bbc1 and
the sample used in the matroska-dovi-write-config8 FATE-test
includes a displaymatrix indicating a rotation which is now
properly written and read, thereby providing coverage for
the relevant code in the muxer as well as the demuxer.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Note: There is a slight difference in the handling of
the max_file_size option: The earlier code used it to mean
to limit the size of the buffer to allocate; the new code
treats it more literally as maximum size to read from
the input.
Reviewed-by: Nicolas George <george@nsup.org>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This avoids including lavc/codec_desc.h everywhere and thereby
forces users to include it directly instead of lazily and potentially
unknowingly relying on indirect inclusions.
Also add the proper inclusion to libavformat/demux.c, one of the
two files that actually use the new field.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The code currently presumes that a return value of AVERROR(ENOMEM)
implies that ac3hdr could not be allocated, so it need not be freed.
Yet any avpriv_ac3_parse_header() might allocate more than the
AC3HeaderInfo internally (it doesn't currently), so simply free
it unconditionally.
Fixes Coverity issues #1492870 and #1492868.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Change some internal APIs a bit to make it harder to make
such mistakes.
In particular, have the read chunk functions return an error
when the result is incomplete.
This might be less flexible, but since there has been no
use-case for that so far, avoiding coding mistakes seems better.
Add a function to queue a AVBPrint directly (ff_subtitles_queue_insert_bprint).
Also fixes a leak in lrcdec when ff_subtitles_queue_insert fails.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
The contents are full TTML XML documents. TTML writing tests'
results are updated as the streams are now properly identified
as TTML ones.
Signed-off-by: Jan Ekström <jan.ekstrom@24i.com>
The size offset was previously being accounted for in flv_set_video_codec
for h264 and mpeg4, instead of being directly accounted for in the spot
where its read, which desynced on HEVC streams.
For clarity, move the size offset directly to the parsing, similar to
how its done for all other header fields.
Fixes: division by zero
Fixes: 57828/clusterfuzz-testcase-minimized-ffmpeg_dem_RKA_fuzzer-6571818338353152
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
I cannot find the spec, but according to the original commit
d4fdba0df7, it's CAVS. e571305a71 changed it to AVS by
accident. Ten years on, nothing happened. We still have the
sample [1], however, since there is no cavs_mp4tofoobar bsf, the
cavs decoder doesn't work. I don't know if there is any use case.
[1] https://samples.ffmpeg.org/AVS/AVSFileFormat/AVSFileFormat.mp4
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
av_random_bytes() can use OS provided strong random functions and does not
depend soley on openssl/gcrypt external libraries.
Fixes ticket #10441.
Signed-off-by: Marton Balint <cus@passwd.hu>
Should set "number of frames" to bytes 24-27 of IVF header, not
duration.
It is described by [1], and confirmed by parsing all IVF files in [2].
This commit also updates the md5sum of refs to pass fate-cbs.
[1] Duck IVF - MultimediaWiki
https://wiki.multimedia.cx/index.php/Duck_IVF
[2] webm/vp8-test-vectors
https://chromium.googlesource.com/webm/vp8-test-vectors
Signed-off-by: Jianhui Dai <jianhui.j.dai@intel.com>
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
Fixes potential use of uninitialized values
in evc_read_nal_unit_length().
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
For uint8_t buf[EVC_NALU_LENGTH_PREFIX_SIZE], &buf still points
to the beginning of buf, but it is of type "pointer to array of
EVC_NALU_LENGTH_PREFIX_SIZE uint8_t" (i.e. pointer arithmetic
would operate on blocks of size EVC_NALU_LENGTH_PREFIX_SIZE).
This is of course a different type than uint8_t*, which is why
there have been casts in evc_read_packet(). But these are unnecessary
if one justs removes the unnecessary address-of operator.
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Content-Type can include charset and boundary which is not a part of
mime type and shouldn't be copied as such.
Fixes HLS playback when the Content-Type includes additional fields.
Signed-off-by: Kacper Michajłow <kasper93@gmail.com>
The discrepancy between the definition and the declaration
in allformats.c is actually UB.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Add demuxer to probe raw vvc and parse vvcc byte stream format.
Co-authored-by: Nuo Mi <nuomi2021@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
This partially reverts commit d0fc1b3507, which reintroduced a regression
originally fixed in 5e9986fd2d.
Signed-off-by: James Almer <jamrial@gmail.com>
mov_try_read_block() allocates 1MB at least, which can be more than
enough. It was called when reading saiz box, which can appear
periodically inside fmp4. This consumes a lot of memory.
We can fix mov_try_read_block() by clamp 'block_size' with 'size'.
However, the function is harmful than helpful. It avoids allocating
large memory when the real data is small. Even in that case, if
allocating large memory directly failed, it's fine to return ENOMEM;
if allocating success and reading doesn't match the given size, it's
fine to free and return AVERROR_INVALIDDATA. In other cases, it's a
waste of CPU and memory.
So I decided to remove the function, and replace it by call
av_malloc() and avio_read() directly.
mov_read_saiz() and mov_read_pssh() need more check, but they don't
belong to this patch.
Fixes#7641 and #9243.
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
But ensure the value returned by evc_read_nal_unit_length() fits in an int.
Should prevent integer overflows later in the code.
Signed-off-by: James Almer <jamrial@gmail.com>
Before the introduction of AV_CODEC_ID_TIMED_ID3 for timed_id3 metadata streams
in mpegts (commit 4a4437c0fb), AV_CODEC_ID_SMPTE_KLV
was the only existing codec for metadata.
It seems that this codec has a 5-bytes metadata header[1] that, for some reason,
was always skipped when decoding data packets.
However, when working with a AV_CODEC_ID_TIMED_ID3 streams, this results in the
5 first bytes of the payload being cut-off, which includes essential informations
such as the ID3 tag version.
This patch fixes the issue by keeping the 5-bytes skip only for AV_CODEC_ID_SMPTE_KLV
streams.
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
Fixes: out of array read
Fixes: 59828/clusterfuzz-testcase-minimized-ffmpeg_dem_JPEGXL_ANIM_fuzzer-5029813220671488
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Incrementing a NULL pointer is undefined behaviour,
yet this is what would happen in case trailer were NULL
before the check.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
In ff_rtp_send_jpeg, the type is defined based on PIX_FMT and
color-range parsed in. There is limitation on current design
where need to include support newly introduced PIX_FMT such as
AV_PIX_FMT_QSV and there might be more and more in future. Hence,
retrive the sampling factor from SOF0 in JPEG compressed header
directly. This introduces flexibility to handle different type of
new codec introduced in future.
Signed-off-by: Yeoh, Hoong Tee <hoong.tee.yeoh@intel.com>
This switches the jpegxl_collect_codestream_header function to use
avcodec/bytestream2, which better enforces barriers, and should avoid
overrunning buffers with jxlp boxes if the size is zero or if the size
is so small the box is invalid.
Signed-off-by: Leo Izen <leo.izen@gmail.com>
- Changes in mov_write_video_tag function to handle EVC elementary stream
- Provided structure EVCDecoderConfigurationRecord that specifies the decoder configuration information for ISO/IEC 23094-1 video content
Signed-off-by: Dawid Kozinski <d.kozinski@samsung.com>
Enable the checked bitreader to avoid overread.
Also add a few checks in loops and between blocks so we exit instead of continued
execution.
Alternatively we could add manual checks so that no overread can happen. This would be
slightly faster but a bit more work and a bit more fragile
Fixes: Out of array accesses
Fixes: 59640/clusterfuzz-testcase-minimized-ffmpeg_dem_JPEGXL_ANIM_fuzzer-6584117345779712
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
The tag comes from samples/ffmpeg/mov/unrecognized/bartjones.mov
really looks like some random data. Now the random tag matched
another file, which isn't a mov.
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
Animated JPEG XL files requires a separate demuxer than image2, because
the timebase information is set by the demuxer. Should the timebase of
an animated JPEG XL file be incompatible with the timebase set by the
image2pipe demuxer (usually 1/25 unless set otherwise), rescaling will
fail. Adding a separate demuxer for animated JPEG XL files allows the
timebase to be set correctly.
Signed-off-by: Leo Izen <leo.izen@gmail.com>
Fixes: signed integer overflow: 9079256848778919936 - -288230376151711746 cannot be represented in type 'long'
Fixes: 58248/clusterfuzz-testcase-minimized-ffmpeg_dem_OGG_fuzzer-6326851353313280
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Check init_get_bits' result for NULL, to avoid dereferencing a NULL
pointer later (CWE-476).
Without this, a segfault happens when trying to decode a handcrafted
ogg-flac file with an absurdly long (e.g. 268435455 bytes) ogg header.
Co-authored-by: James Almer <jamrial@gmail.com>
Signed-off-by: Paul Arzelier <paul.arzelier@free.fr>
Fixes: signed integer overflow: -9223372036854775808 - 2082844800 cannot be represented in type 'long'
Fixes: 58384/clusterfuzz-testcase-minimized-ffmpeg_dem_MOV_fuzzer-6428383700713472
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
This should be a few nano seconds faster (not measureable)
But Collectively the whole humankind watching hls will safe a minute
Found-by: Leo Izen
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
The demuxer exports rawvideo, so there's no reason for the muxer to only
work with wrapped_avframe.
Reviewed-by: Ronald S. Bultje <rsbultje@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
For encoding, this field is entirely redundant with
AVCodecContext.framerate.
For decoding, this field is entirely redundant with
AV_CODEC_PROP_FIELDS.
H.264 and mpeg12 parsers need to be adjusted at the same time to stop
using the value of AVCodecContext.ticks_per_frame, because it is not set
correctly unless the codec has been opened. Previously this would result
in both the parser and lavf seeing the same incorrect value, which would
cancel out.
Updating lavf and not the parsers would result in correct value in lavf,
but the wrong one in parsers, which would break some tests.
This should fix the regression since 6b1f68ccb0
Should fix Ticket10353 (please test and report cases that still fail)
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
This patch simply recognizes the AAC audio track during
decode -- it does not add functionality to encode AAC in
MXF.
Signed-off-by: Ammon Riley <ammon.riley@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>
Its unexpected that a .avi or other "standard" file turns into a playlist.
The goal of this patch is to avoid this unexpected behavior and possible
privacy or security differences.
Reviewed-by: Steven Liu <lingjiujianke@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
One that is fine enough to represent all DV audio sample rates. Audio
packet durations are now sample-accurate.
This largely undoes commit 76fbb0052d. To
avoid breaking the issue fixed by that commit, resync audio timestamps
against video if they get more than one frame apart. The sample from
issue #8762 still works correctly after this commit.
Slightly changes the results of the lavf-dv seektest, due to the audio
timebase being more granular.
Current code will call avpriv_set_pts_info() for each video frame,
possibly setting a different timebase if the stream framerate changes.
This violates API conventions, as the timebase is supposed to stay
constant after stream creation.
Change the demuxer to set a single timebase that is fine enough to
handle all supported DV framerates.
The seek tests change slightly because the new timebase is more
granular.
This was a bug/mistake in dae3679a9b.
use_mfra_for by defintion only has an effect on fragmented MP4 files,
making the check not only redundant, but also broken if a user used
the option globally (i.e. set on non-fragmented MP4s).
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Also make duration check for mvhd more consistent with the others, write
version 1 of mvhd if duration is at least INT32_MAX instead of UINT32_MAX.
Signed-off-by: Marton Balint <cus@passwd.hu>
Commit 23eeffcd48 added a hack to support invalid
files where the creation date was encoded as a classic unix timestamp. Let's
reduce the scope of the hack by only applying it to version 0 mdhd/mvhd atoms.
Also warn the user of such possibly broken files.
Signed-off-by: Marton Balint <cus@passwd.hu>
When write multi-trun box, the MOV_TRUN_FIRST_SAMPLE_FLAGS flag
need judge by first param, not 0.
If the original video contains consecutive I frames,
this will cause the packets of fmp4 have error sample_flags ,
and then incorrect keyframes were generated,
and then error packet will be seeked.
Signed-off-by: Wang Yaqiang <wangyaqiang03@kuaishou.com>
Signed-off-by: Steven Liu <liuqi05@chinaffmpeg.org>
Changes the result of the h264_redundant_pps-mov test, where the output
timebase is now 1001/24000 instead of 1/24. This is more correct, as the
source file actually is 23.98fps.
Timestamps in two FATE H.264 conformance tests now start at 1 instead
of 0, which also happens in some other H.264 tests before this commit
and so is not a big issue.
Conversely, timestamps in some HEVC conformance tests start from a
smaller value now.
Ideally this should be addressed later in a more general way.
h264-conformance-frext-frext2_panasonic_b no longer requires -vsync
passthrough.
The Matroska spec requires it to be equal to the highest BlockAddID value in a
BlockAdditions, but being purely an informative element, only abort if strict
compliance is requested, as demuxing is otherwise unaffected.
Signed-off-by: James Almer <jamrial@gmail.com>
RIP, if exists is the last KLV item in the MXF files therefore we can stop
parsing the file if it is encountered. This allows us to support files created by
broken muxers such as OpenCube MXFTk Advanced 2.8.0.0.1. which dumps some extra
garbage after the RIP.
Signed-off-by: Marton Balint <cus@passwd.hu>
Current log messages talk about 'suitable' output formats. This is
confusing, because it suggests that some formats are suitable, while
others are not, which is not the case.
Also, suggest possible user actions when format cannot be guessed from a
filename.
An uninitialized AVFormatContext instance with neither iformat nor
oformat set will currently log as 'NULL', which is confusing and
unhelpful. Print 'AVFormatContext' instead, which provides more
information.
This happens e.g. if choosing an output format fails in
avformat_alloc_output_context2().
E.g. with the following commandline:
ffmpeg -i <input> -f foobar -y /dev/null
before: [NULL @ 0x5580377834c0] Requested output format 'foobar' is not a suitable output format
after: [AVFormatContext @ 0x55fa15bb34c0] Requested output format 'foobar' is not a suitable output format
3GPP TS 26.244 Table 8.10 specifies that longitude is written before
latitude. The MOV demuxer already expects the correct order. So, write
them in that order.
However, the user supplied string may also be used in MOV mode which
requires ISO 6709 format which specifies latitude first. The demuxer
also exports the loci value in that format. So parser adjusted as well.
When writing a subtitle SSA/ASS subtitle file, the AVCodecParameters::extradata
buffer is written directly to the output. In the case where the buffer is
filled from a matroska source file produced by some older versions of
Handbrake, this buffer ends with a null terminating character, which is then
erroneously copied into the middle of the output file. The change here avoids
this problem by treating it as a string rather than a raw buffer. This way it
is agnostic as to whether the source buffer was null terminated or not.
Fixes ticket #10203.
Reported-by: Tim Angus <tim at ngus.net>
Signed-off-by: Marton Balint <cus@passwd.hu>
Previously, the ff_configure_buffers_for_index function had
upper sanity limits of 16 MB (1<<24) for buffer_size and
8 MB (1<<23) for short_seek_threshold.
However, if the index contained entries with a much larger
delta, setting pos_delta to a value larger than the sanity
limit, we would end up not increasing the buffer size at all.
Instead, ignore the individual deltas that are excessive, but
increase the buffer size based on the deltas that are below the
sanity limit.
Only count deltas that are below 1<<23, 8 MB; pos_delta gets doubled
before setting the buffer size - this matches the previous maximum
buffer size of 1<<24, 16 MB.
This can happen e.g. with a mov file with some tracks containing
some samples that belong in the start of the file, at the end of
the mdat, while the rest of the file is mostly reasonably interleaved;
previously those samples caused the maximum pos_delta to skyrocket,
skipping any buffer size enlargement.
Signed-off-by: Martin Storsjö <martin@martin.st>
When scanning through the index, account for the fact that the
compared samples may be located in an unexpected order in the file;
this function is mainly interested in the absolute difference between
file locations.
Signed-off-by: Martin Storsjö <martin@martin.st>
Some additional properties are set for ARIB caption.
* need_parsing = 0
ARIB caption doesn't require any parser.
This avoids "parser not found" warning message.
* need_context_update = 1
When any profiles are changed, set this flag to notify.
Signed-off-by: rcombs <rcombs@rcombs.me>
According to MXF specs the Stored Rectangle corresponds to the data which is
passed to the compressor and received from the decompressor, so they should
contain the width / height extended to the macroblock boundary.
In practice however width and height values rounded to the upper 16 multiples
are only seen when muxing MPEG formats. Therefore this patch changes stored
width and height values to unrounded for all non-MPEG formats, even macroblock
based ones.
For DNXHD the specs (ST 2019-4) explicitly indicates to use 1080 for 1088p.
For ProRes the specs (RDD 44) only refer to to ST 377-1 without precision but
no known commercial implementations are using rounded values.
DV is not using 16x16 macroblocks, so 16 rounding makes no sense.
The patch also fixes Sampled Width / Display Width to use unrounded values.
Signed-off-by: Marton Balint <cus@passwd.hu>
Add the appropriate descriptors to the MPEG-TS demux and mux to
ensure that SMPTE 2038 VANC streams are properly preserved
when using codec copy (including adding the appropriate PMT
descriptors).
The focus of this patch is TS input to TS output. A separate
patch adds support for output of 2038 VANC over decklink SDI.
Thanks to Marton Balint for feedback to preserve ABI compatibility.
Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com>
Signed-off-by: Marton Balint <cus@passwd.hu>
The original code would strip off the PTS/DTS of any packets
which had a stream ID of STREAM_ID_PRIVATE_STREAM_1. While the
intent was to apply this to asynchronous KLV packets, it was
being applied to any codec that had that same stream ID (including
types such as SMPTE 2038).
Add a clause to the if() statement to ensure it only gets applied
if the codec actually is KLV.
Signed-off-by: Devin Heitmueller <dheitmueller@ltnglobal.com>
Signed-off-by: Marton Balint <cus@passwd.hu>
ipcm is defined by ISO/IEC 23003-5, not defined by quicktime. After
adding ISO/IEC 23003-5 support, we don't need it for ticket #9219.
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
svg is xml, but <?xml is not required,
it can start with <svg and can have multiple empty lines,
or start with <!-- include some comments,
but must first line if start with <?xml.
Signed-off-by: Wang Yaqiang <wangyaqiang03@kuaishou.com>
Signed-off-by: Marton Balint <cus@passwd.hu>
The path attribute in the Set-Cookie header is optional but treated by ffmpeg as being compulsory.
Signed-off-by: Michael J. Walsh <mjfwalsh@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>
As per 23003-5:2020, the rest of the bits are reserved, and thus
in the future they may be utilized for something else.
Quote:
format_flags is a field of flags that modify the default PCM sample format.
Undefined flags are reserved and shall be zero. The following flag is defined:
0x01 indicates little-endian format. If not present, big-endian format is used.
When hls_init_time should available when hls_list_size > 0.
Because the list will not refresh new top line segment when hls_list_size is 0
Signed-off-by: Steven Liu <lq@chinaffmpeg.org>
Fixes: division by zero
Fixes: 55940/clusterfuzz-testcase-minimized-ffmpeg_IO_DEMUXER_fuzzer-6333107679920128
The decoder does not support bps=1 and i have no such sample so it is not
known if this duration is correct. Alternatively we could error out on all
bps we currently do not support on the decoder side or not set duration.
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Fixes: signed integer overflow: 9223372036854775584 + 536870912 cannot be represented in type 'long'
Fixes: 55844/clusterfuzz-testcase-minimized-ffmpeg_dem_MOV_fuzzer-510613920664780
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
Current mpegtsenc code only inserts SPS/PPS from extradata before IDR frames if
AUD is also inserted.
Unfortunately some encoders may preface a key frame with an AUD, but no
SPS/PPS. In that case current code does not repeat the "extradata" and the
resulting HLS stream may become noncompliant and unjoinable.
Fix this by always inserting SPS/PPS and moving AUD to the beginning of the
packet if it is already present.
Fixes ticket #10148.
Signed-off-by: Marton Balint <cus@passwd.hu>
FLV spec only has AVC end of sequence tag, and the EOS tag has a
CodecID as other video data packet. MPEG4 doesn't conformance to
the spec, but it's there for a decade. So only 'fix' the EOS tag
rather than remove it completely.
Reviewed-by: Steven Liu <lq@chinaffmpeg.org>
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
