Such files have IndexTableSegments which when parsed cover EditUnit
ranges like this:
[0,1)
[249,250)
[249,377)
[0,249)
where each interval is
[IndexStartPosition, IndexStartPosition + IndexDuration)
This would be reduced to a sparse index like:
[0,1), [249,250)
instead of the full range:
[0,249), [249,377)
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Since 83cab07 audio stream time bases are based on SampleRate, not EditRate.
This fixes trac ticket #2029 and a few seeking issues.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
None of these are likely unless the user is writing a file with two billion
streams or a duration of around two months.
This fixes CIDs 700568, 700569, 700570, 700571, 700572 and 700573.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
There's a a potential DoS problem in this function. Say an MXF file is
created with a PixelLayout with a long run of non-zeroes. Such a file could be
sent quickly (packed) over the net and would unpack quite fast. mxfdec would
then read it byte-by-byte, which would take considerable time.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Coverity thinks ofs can end up 15, thus writing past the end of layout[]. This
is incorrect since it's always incremented by 2. Checking ofs <= 14 makes
Coverity happy and doesn't hurt.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
We cannot do this in general since we could be reading
a file with B-frames while lacking an index.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
This allows future assumptions to be made without affecting non-intra files.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
The "ECs != 1 -> OP1a" assumption was wrong. Luckily, the file
that triggered that behavior had two ECs, not zero. Hence
distinguishing between them is simple in this case.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
This fixes rare cases where OPAtom may be treated
as OP1a, causing all essence to be read into RAM.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
The reason for this is that such files have IndexTableSegments which when parsed
cover EditUnit ranges like this:
[0,1)
[249,250)
[249,377)
[0,249)
where each interval is [IndexStartPosition,IndexStartPosition+IndexDuration).
This would be reduced to a sparse index like:
[0,1), [249,250)
instead of the full range:
[0,249), [249,377)
See TimeCode_HD.mxf, UMID =
060a2b340101010101010410130000000004001aa0e59175025b2a5600da4101.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
We can't do this in general since we could be reading a file with B-frames while
lacking an index.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The "ECs != 1 -> OP1a" assumption was wrong. Luckily, the file that triggered
that behavior had two ECs, not zero. Hence distinguishing between them is
simple in this case.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This fixes rare cases where OPAtom may be treated as OP1a, causing all essence
to be read into RAM.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
These are common with audio atoms. Without this the demuxer would read two
bytes at a time for a mono 16-bit file.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
Only the OPAtom demuxing logic is guaranteed to have index tables,
meaning OP1a files that lack an index would cause SIGSEGV.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
Specifically, this means parsing as before until we run into essence.
At that point we seek to the footer and parse until EOF. After that we start
seeking backward to the previous partition and parse that until we run into
essence or the next partition. This procedure is repeated until we encounter
the last partition we parsed in the forward direction.
The end result of all this is that large essence containers are not needlessly
parsed. This speeds up parsing large files a lot.
Signed-off-by: Janne Grunau <janne-libav@jannau.net>
This fixes 0001GL.MXF.V1.mxf_opatom.mxf and 0001GL00.MXF.A1.mxf_opatom.mxf
getting two streams each due to both using the same SourcePackageID.
Signed-off-by: Janne Grunau <janne-libav@jannau.net>
Only the OPAtom demuxing logic is guaranteed to have index tables, meaning OP1a
files that lack an index would cause SIGSEGV.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This uses the old demuxing code for OP1a and separate demuxing code for OPAtom.
Timestamp output is added to the old demuxing code.
The seeking code is made to seek to the start of the desired EditUnit only,
from which the normal demuxing code takes over (if OP1a). This means we don't
use delta entries or slices, only StreamOffsets.
OPAtom seeking basically works like before.
This also makes D-10 seeking behave the same way as OP1a and OPAtom. In other
words, we allow seeking before the start or past the end for D-10 too.
This fixes ticket #746.
This changes mxf_compute_ptses() to be used for MXFIndexTable, and also adds
code for computing the fake index to it.
This also temporarily disables PTS computation. A future patch will restore it.
These indexes duplicate every entry and have the total size of the essence
container as the last entry.
This patch also computes the size of the packets when unknown.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
I thought it had to do with file offsets, but's actually the offset inside
the essence container.
In other words, unbreak multiple EditUnitByteCounts.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
These values include KAGSize, HeaderByteCount and IndexByteCount.
The length of the pack itself is also stored, and KAGSize is sanity checked.
The FATE sample has KAGSize == 0, which is adjusted to 512.
Other bad KAGSizes are set to 1.
Specifically, this means parsing as before until we run into essence.
At that point we seek to the footer and parse until EOF. After that we start
seeking backward to the previous partition and parse that until we run into
essence or the next partition. This procedure is repeated until we encounter
the last partition we parsed in the forward direction.
The end result of all this is that large essence containers aren't needlessly
parsed. This speeds up parsing large files a lot.
Extradata should only be parsed from the avss, fiel, jp2h and alac atoms for
AVS, MJPEG, Motion JPEG 2000 and ALAC respectively.
This also fixes the mov demuxer coming up with bogus extradata for some
AVC-Intra samples due to the presence of fiel atoms.
Moving the search and parsing of the 'fmt ' info the main loop of
wav_read_header() allows tags that precede it to be parsed. Creating
wav_parse_fmt_tag() makes wav_read_header() easier to read.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
Set DV packet durations using fields_per_frame.
This requires turning gxf_stream_info into the demuxer's context for access to the value in gxf_packet().
Since MPEG-2 seems to work fine this done only for DV.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
Moving the search and parsing of the 'fmt ' info the main loop of wav_read_header() allows tags that precede it to be parsed.
Creating wav_parse_fmt_tag() makes wav_read_header() easier to read.
This fixes PCR drift due to accumulating TS_PACKET_SIZE*8*90000LL/ts->mux_rate each packet, due to rounding errors when mux_rate does not evenly divide 135360000.
This patch also increases the PCR precision to 27 MHz from 90 kHz and takes the location of the PCR data into account (+11 bytes according to the spec).
Originally committed as revision 25864 to svn://svn.ffmpeg.org/ffmpeg/trunk
rawdec.c: Only perform bits_per_coded_sample -> pix_fmt guesswork if pix_fmt hasn't been set
Originally committed as revision 23561 to svn://svn.ffmpeg.org/ffmpeg/trunk
