GPU copy enables or disables GPU accelerated copying between video
and system memory. This may lead to a notable performance improvement.
Memory must be sequent and aligned with 128x64.
CMD:
ffmpeg -init_hw_device qsv=hw -filter_hw_device hw -c:v h264_qsv
-gpu_copy on -i input.h264 -f null -
or:
ffmpeg -c:v h264_qsv -gpu_copy on -i input.h264 -f null -
Signed-off-by: Linjie Fu <linjie.fu@intel.com>
Signed-off-by: ChaoX A Liu <chaox.a.liu@intel.com>
Signed-off-by: Zhong Li <zhong.li@intel.com>
There is no change in the encoded bitstream, but this
ensures that the written field length is consistent
with the reference implementation.
Unused bytes are zeroed out for backwards compatibility.
Signed-off-by: Raphaël Zumer <rzumer@tebako.net>
the data_start is after reading 12 bytes and if its subtracted
at the very end the intermediate might overflow
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
This seeks to the position the previous call to dxv_decompress_opcodes()
positioned us in case of success
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
made with persistent connections to prevent incorrect reset of
offset when demuxing HLS+FMP4
Signed-off-by: Steven Liu <lq@onvideo.cn>
Signed-off-by: vectronic <hello.vectronic@gmail.com>
add ff_http_do_new_request2() which supports options to be applied to
HTTPContext after initialisation with the new uri
Signed-off-by: Steven Liu <lq@onvideo.cn>
Signed-off-by: vectronic <hello.vectronic@gmail.com>
Linux and OSX systems support basename and dirname via <libgen.h>, I plan to
make the wrapper interface conform to the standard interface first.
If it is feasible, I will continue to modify it to call the system interface
if there is already a system call interface.
You can get more description about the system interface by below command:
"man 3 basename"
Reviewed-by: Marton Balint <cus@passwd.hu>
Reviewed-by: Tomas Härdin <tjoppen@acc.umu.se>
Reviewed-by: Steven Liu <lq@chinaffmpeg.org>
Signed-off-by: Limin Wang <lance.lmwang@gmail.com>
Current default is 200kbps, which produces inconsistent
results (too high for low-res, too low for hi-res). Use
CRF instead, which will adapt. Affects VP9. Also have
VP8 use a default bitrate of 256kbps.
Signed-off-by: James Zern <jzern@google.com>
When flushing, MAX_FRAME_HEADER_SIZE bytes (always zero) are supposed to
be written to the fifo buffer in order to be able to check the rest of
the buffer for frame headers. It was intended to write these by writing
a small buffer of size MAX_FRAME_HEADER_SIZE to the buffer. But the way
it was actually done ensured that this did not happen:
First, it would be checked whether the size of the input buffer was zero,
in which case it buf_size would be set to MAX_FRAME_HEADER_SIZE and
read_end would be set to indicate that MAX_FRAME_HEADER_SIZE bytes need
to be written. Then it would be made sure that there is enough space in
the fifo for the data to be written. Afterwards the data is written. The
check used here is for whether buf_size is zero or not. But if it was
zero initially, it is MAX_FRAME_HEADER_SIZE now, so that not the
designated buffer for writing MAX_FRAME_HEADER_SIZE is written; instead
the padded buffer (from the stack of av_parser_parse2()) is used. This
works because AV_INPUT_BUFFER_PADDING_SIZE >= MAX_FRAME_HEADER_SIZE.
Lateron, buf_size is set to zero again.
Given that since 7edbd536, the actual amount of data read is no longer
automatically equal to buf_size, it is completely unnecessary to modify
buf_size at all. Moreover, modifying it is dangerous: Some allocations
can fail and because buf_size is never reset to zero in this codepath,
the parser might return a value > 0 on flushing.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
For a parser, the input buffer is always != NULL: In case of flushing,
the indicated size of the input buffer will be zero and the input buffer
will point to a zeroed buffer of size 0 + AV_INPUT_BUFFER_PADDING.
Therefore one does not need to check for whether said buffer is NULL or
not.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
The flac parser uses a fifo to buffer its data. Consequently, when
searching for sync codes of flac packets, one needs to take care of
the possibility of wraparound. This is done by using an optimized start
code search that works on each of the continuous buffers separately and
by explicitly checking whether the last pre-wrap byte and the first
post-wrap byte constitute a valid sync code.
Moreover, the last MAX_FRAME_HEADER_SIZE - 1 bytes ought not to be searched
for (the start of) a sync code because a header that might be found in this
region might not be completely available. These bytes ought to be searched
lateron when more data is available or when flushing.
Unfortunately there was an off-by-one error in the calculation of the
length to search of the post-wrap buffer: It was too large, because the
calculation was based on the amount of bytes available in the fifo from
the last pre-wrap byte onwards. This meant that a header might be
parsed twice (once prematurely and once regularly when more data is
available); it could also mean that an invalid header will be treated as
valid (namely if the length of said invalid header is
MAX_FRAME_HEADER_SIZE and the invalid byte that will be treated as the
last byte of this potential header happens to be the right CRC-8).
Should a header be parsed twice, the second instance will be the best child
of the first instance; the first instance's score will be
FLAC_HEADER_BASE_SCORE - FLAC_HEADER_CHANGED_PENALTY ( = 3) higher than
the second instance's score. So the frame belonging to the first
instance will be output and it will be done as a zero length frame (the
difference of the header's offset and the child's offset). This has
serious consequences when flushing, as returning a zero length buffer
signals to the caller that no more data will be output; consequently the
last frames not yet output will be dropped.
Furthermore, a "sample/frame number mismatch in adjacent frames" warning
got output when returning the zero-length frame belonging to the first
header, because the child's sample/frame number of course didn't match
the expected sample frame/number given its parent.
filter/hdcd-mix.flac from the FATE-suite was affected by this (the last
frame was omitted) which is the reason why several FATE-tests needed to
be updated.
Fixes ticket #5937.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
The FLACHeaderMarker structure contained a pointer to an array of int;
said array was always allocated and freed at the same time as its
referencing FLACHeaderMarker; the pointer was never modified to point to
a different array and each FLACHeaderMarker had its own unique array.
Furthermore, all these arrays had a constant size. Therefore include
this array in the FLACHeaderMarker struct.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
FLAC sync codes contain a byte equal to 0xFF and so the function that
searches for sync codes first searched for this byte. It did this by
checking four bytes at once; these bytes have been read via AV_RB32, but
the test works just as well with native endianness.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Linjie Fu
Zhong Li
Raphaël Zumer
Zhao Zhili
Michael Niedermayer
Paul B Mahol
Andreas Rheinhardt
vectronic
Limin Wang
Steven Liu
elliottk