It is my understanding that "Unless otherwise stated, all data in a
QuickTime movie is stored in big-endian byte ordering" [1] in MOV files.
I have a couple of thousand files, which technically are invalid because
their sound sample description element 4CC is 'lpcm' but its version is
0 - and "Version 0 supports only uncompressed audio in raw ('raw ') or
twos-complement ('twos') format" [2]
Because isom.c only contains a mapping for 4CC 'lpcm' to
AV_CODEC_ID_PCM_S16LE, these files have their audio decoded as LE when
it is actually BE.
This commit adds AV_CODEC_ID_PCM_S16BE as the first match for 4CC 'lpcm'.
[1]
https://developer.apple.com/library/mac/documentation/quicktime/QTFF/qtff.pdf
page 21
[2]
https://developer.apple.com/library/mac/documentation/quicktime/QTFF/qtff.pdf
page 178
Reviewed-by: Yusuke Nakamura <muken.the.vfrmaniac@gmail.com>
Not guaranteed to be in nanosecond resolution. On iOS 7 the duration
of one tick is 125/3 ns which is still more than an order of magnitude
better then microseconds.
Replace decicycles with the neutral UNITS. Decicycles is strange but
tenths of a nanosecond and unspecific "deci"-ticks for mach_absolute_time
is just silly.
Based on the aarch64 asm. CPU cycle counts on cortex-a9 compared to
gcc 4.8.2:
before: 475 decicycles in get_cabac_noinline, 67106035 runs, 2829 skips
after: 393 decicycles in get_cabac_noinline, 67106474 runs, 2390 skips
Overall speedup is above 2%. Code generated by clang 3.4 is slower on
the same hardware and the relative change is a little larger.
Based on the x86 branchless get_cabac asm. get_cabac_noinline() gets
approximately 20% faster (no cycle counts available) compared to clang
from Xcode 5.1 beta5. More than 6% faster overall. A part of the overall
speedup might be explained by additional inlining of get_cabac().
The overread avoidance fix in cbddee1cca
broke the computation for the last row since it prevented the safe
reading from the height+1-th row.
CC: libav-stable@libav.org
Based on a suggestion by Martin Panter. This is more descriptive,
since it's the actual timestamp field from the RTMP packet,
which might or might not be a delta depending on context (in
some packets it's a delta, in some packets it's an absolute
timestamp, and in some packets it's 0xffffff to indicate that
the actual delta or absolute timestamp is transmitted separately).
Signed-off-by: Martin Storsjö <martin@martin.st>
Related fix in "rtmpdump":
https://repo.or.cz/w/rtmpdump.git/commitdiff/79459a2
Adobe's RTMP specification (21 Dec 2012), section 5.3.1.3 ("Extended
Timestamp"), says "this field is present in Type 3 chunks". Type 3 chunks are
those with the one-byte header size.
This resolves intermittent hangs and segfaults caused by the read function,
and also includes an untested fix for the write function.
The read function was tested with ABC (Australia) News 24 streams, however
they are probably restricted to only Australian internet addresses. Some of
the packets at the start of these streams seem to contain junk timestamp
fields, often requiring the extended field. Test command:
avplay rtmp://cp81899.live.edgefcs.net/live/news24-med@28772
Signed-off-by: Martin Storsjö <martin@martin.st>
Some content requires an higher number of slices in order to
render properly.
Rise the number to 1024 and warn if ever it exceeds.
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Get the last partition offset and use it when footer partition
offset is missing.
Footer partition may not be present and even if present footer
partition offset may not be set in any partition except last one.
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Extrapolate audio timestamps based on the number of samples demuxed.
Deal with some MXF nastiness involving fractional number of
samples per EditUnit when seeking (the specs handwave this away).
Further fixes from Tomas Härdin.
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
The SVQ3 decoder reuses large parts of the H.264 decoder so it
makes no sense to enable the former but not the latter.
Also drop unnecessary h263.o object from SVQ3 decoder object list.
The s390 architecture requires shared libraries to be built in PIC mode.
Otherwise applications will get wrong relocations at run-time, leading
to confusing segmentation faults.
CC: libav-stable@libav.org
