Commit Graph

23 Commits

Author SHA1 Message Date
Ben Avison
701e8b42e1 vc-1: Optimise parser (with special attention to ARM)
The previous implementation of the parser made four passes over each input
buffer (reduced to two if the container format already guaranteed the input
buffer corresponded to frames, such as with MKV). But these buffers are
often 200K in size, certainly enough to flush the data out of L1 cache, and
for many CPUs, all the way out to main memory. The passes were:

1) locate frame boundaries (not needed for MKV etc)
2) copy the data into a contiguous block (not needed for MKV etc)
3) locate the start codes within each frame
4) unescape the data between start codes

After this, the unescaped data was parsed to extract certain header fields,
but because the unescape operation was so large, this was usually also
effectively operating on uncached memory. Most of the unescaped data was
simply thrown away and never processed further. Only step 2 - because it
used memcpy - was using prefetch, making things even worse.

This patch reorganises these steps so that, aside from the copying, the
operations are performed in parallel, maximising cache utilisation. No more
than the worst-case number of bytes needed for header parsing is unescaped.
Most of the data is, in practice, only read in order to search for a start
code, for which optimised implementations already existed in the H264 codec
(notably the ARM version uses prefetch, so we end up doing both remaining
passes at maximum speed). For MKV files, we know when we've found the last
start code of interest in a given frame, so we are able to avoid doing even
that one remaining pass for most of the buffer.

In some use-cases (such as the Raspberry Pi) video decode is handled by the
GPU, but the entire elementary stream is still fed through the parser to
pick out certain elements of the header which are necessary to manage the
decode process. As you might expect, in these cases, the performance of the
parser is significant.

To measure parser performance, I used the same VC-1 elementary stream in
either an MPEG-2 transport stream or a MKV file, and fed it through avconv
with -c:v copy -c:a copy -f null. These are the gperftools counts for
those streams, both filtered to only include vc1_parse() and its callees,
and unfiltered (to include the whole binary). Lower numbers are better:

                Before          After
File  Filtered  Mean   StdDev   Mean   StdDev  Confidence  Change
M2TS  No        861.7  8.2      650.5  8.1     100.0%      +32.5%
MKV   No        868.9  7.4      731.7  9.0     100.0%      +18.8%
M2TS  Yes       250.0  11.2     27.2   3.4     100.0%      +817.9%
MKV   Yes       149.0  12.8     1.7    0.8     100.0%      +8526.3%

Yes, that last case shows vc1_parse() running 86 times faster! The M2TS
case does show a larger absolute improvement though, since it was worse
to begin with.

This patch has been tested with the FATE suite (albeit on x86 for speed).

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
2014-08-04 22:22:54 +02:00
Diego Biurrun
6fee1b90ce avcodec: Add av_cold attributes to init functions missing them 2013-05-04 21:09:45 +02:00
Masaki Tanaka
accde1bd87 vc1_parser: Set field_order.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
2013-05-03 08:20:19 +02:00
Anton Khirnov
36ef5369ee Replace all CODEC_ID_* with AV_CODEC_ID_* 2012-08-07 16:00:24 +02:00
Ronald S. Bultje
c742ab4e81 vc1parse: call vc1_init_common().
The parser uses VLC tables initialized in vc1_common_init(), therefore
we should call this function on parser init also.

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC: libav-stable@libav.org
2012-02-18 09:16:39 -08:00
Martin Storsjö
5f2c159c0f vc1: Add ff_ prefix to nonstatic symbols
Signed-off-by: Martin Storsjö <martin@martin.st>
2012-02-15 22:07:05 +02:00
Rafaël Carré
797639dcfd vc1: use ff_parse_close
It works as long as ParseContext is the first member of the private struct

Signed-off-by: Diego Biurrun <diego@biurrun.de>
2012-02-10 15:48:52 +01:00
Janne Grunau
f907615f08 parsers: initialize MpegEncContext.slice_context_count to 1
The mpeg4 video, H264 and VC-1 parser hold (directly or indirectly)
a MpegEncContext in their private context. Since they do not call the
common mpegvideo init function slice_context_count has explicitly set
to 1.
Prevents a null pointer dereference in the h264 parser and fixes
bug 193.
2012-01-06 01:47:45 +01:00
Diego Biurrun
58c42af722 doxygen: misc consistency, spelling and wording fixes 2011-12-12 23:06:23 +01:00
Anton Khirnov
5511ad14fe lavc: use designated initialisers for parsers. 2011-11-02 10:03:43 +01:00
John Stebbins
0d802ac54e vc1: fix VC-1 Pulldown handling.
Pulldown flags are being set incorrectly and AVFrame->repeat_pict is not
being set.  Also, skipped frames exit header parsing too early and do not
set pulldown flags appropriately. Ticks_per_frame needs to be set and
time_base adjusted so player can extend frame duration by a field time.

This fixes problems encountered when attempting to transcode HD-DVD EVOB
files with HandBrake. Also makes these files play smoothly in avplay.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
2011-08-25 12:41:45 -07:00
Stefano Sabatini
975a1447f7 Replace deprecated FF_*_TYPE symbols with AV_PICTURE_TYPE_*.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
2011-05-02 12:18:44 +02:00
Mans Rullgard
2912e87a6c Replace FFmpeg with Libav in licence headers
Signed-off-by: Mans Rullgard <mans@mansr.com>
2011-03-19 13:33:20 +00:00
Diego Elio Pettenò
d36beb3f69 Add ff_ prefix to data symbols of encoders, decoders, hwaccel, parsers, bsf.
None of these symbols should be accessed directly, so declare them as
hidden.

Signed-off-by: Mans Rullgard <mans@mansr.com>
2011-01-26 16:08:45 +00:00
Diego Biurrun
ba87f0801d Remove explicit filename from Doxygen @file commands.
Passing an explicit filename to this command is only necessary if the
documentation in the @file block refers to a file different from the
one the block resides in.

Originally committed as revision 22921 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-04-20 14:45:34 +00:00
Baptiste Coudurier
2d05bc8641 set pict_type in VC-1 parser, fix some timestamps problems
Originally committed as revision 18987 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-05-30 00:09:00 +00:00
Michael Niedermayer
7bbc686f82 Fix vc1 split().
Fixes Subtitle-sample.evo, issue52.

Originally committed as revision 17533 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-02-22 20:48:12 +00:00
Diego Biurrun
bad5537e2c Use full internal pathname in doxygen @file directives.
Otherwise doxygen complains about ambiguous filenames when files exist
under the same name in different subdirectories.

Originally committed as revision 16912 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-02-01 02:00:19 +00:00
Aurelien Jacobs
c53d2d9042 make some parser parameters const to avoid casting const to non-const
Originally committed as revision 8921 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-05-07 00:47:03 +00:00
Kostya Shishkov
de53b04b53 Make vc1_parser.c compilable without special defines
Originally committed as revision 8914 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-05-06 11:01:25 +00:00
Diego Biurrun
bbb7d4c76b Remove superfluous #includes, parser.h now includes its prerequisites.
Originally committed as revision 8905 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-05-05 21:11:22 +00:00
Kostya Shishkov
58683d2743 100l to myself. Do not include stuff unneeded by parser
Originally committed as revision 8886 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-05-04 12:31:40 +00:00
Diego Biurrun
5982ae94ee Move VC1 parser to its own file.
Originally committed as revision 8882 to svn://svn.ffmpeg.org/ffmpeg/trunk
2007-05-04 00:09:33 +00:00