The code to initialize the ungrouped bap mantissa tables
(bap 3 or 5) takes more bytes of .text than the tables itself;
they have therefore been hardcoded.
For GCC (14, -O3, albeit in an av_cold function), the initialization
code takes 99B each for the fixed and floating point decoders
(the code is currently duplicated), whereas the hardcoded tables
only take 96B. For Clang 19 it were 374B each (I don't now what
Clang was doing there).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Setting print_graphs_format (in case no -print_graphs_format
option was specified) is racy, as is using av_strtok()
to split it. Both have been removed.
Notice that using av_strtok() was destructive: In the absence
of races the options would only have been applied for the
first initialization.
Reviewed-by: softworkz . <softworkz-at-hotmail.com@ffmpeg.org>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
As the names imply, they are structs meant to be internal and private to the
filter handling code. If a field is required in other modules, then it can
be moved to the public facing structs, which is done in this commit.
Signed-off-by: James Almer <jamrial@gmail.com>
As the name implies, it's a struct meant to be internal and private to the
filter handling code. If a field is required in other modules, then it can
be moved to the public facing struct, which is done in this commit.
Signed-off-by: James Almer <jamrial@gmail.com>
The follow cmd output corrupted file before the patch:
ffmpeg -f lavfi -i color=blue,trim=duration=0.04 \
-f lavfi -i anullsrc,atrim=duration=2 \
-movflags +empty_moov+hybrid_fragmented \
-frag_duration 1000000 \
-frag_interleave 1 \
output.mp4
1. first_track is the first track with track->entry != 0. As in the
command above, video track (track index 0) has a single frame. When
flush the second fragment, first_track is 1, the audio track.
2. write_moof = i == first_track, so write_moof is false for i = 0.
3. When mov->frag_interleave != 0, mov->mdat_buf != NULL, because
it contains audio data. So avio_write is called before write_moof,
that is, the data write before moof, and mov_finish_fragment
executed with wrong mdat_start.
4. With normal fmp4 output, the error isn't obvious. With
hybrid_fragmented, ffplay output.mp4 shows a lot of error messages.
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
The default behavior for VVenC (since v1.10.0) is to create an IDR with
leading pictures for the first picture in decoding order (POC 32). This
leads to FFmpeg generating an edit list with an empty entry, skipping
the leading pictures.
This patch fixes the calculation for the start_pts, while the DTS is
negative (as produced by vvenc).
Signed-off-by: Gabriel Hege <g+ffmpeg@hege.cc>
The MMX requantize functions have the MMX permutation
(i.e. FF_IDCT_PERM_SIMPLE) hardcoded and therefore
check for the used permutation (namely via a CRC).
Yet this is very ugly and could even lead to misdetection;
furthermore, since d7246ea9f2
the permutation used here is de-facto and since
bfb28b5ce8 definitely
impossible on x64, making this code dead on x64.
So remove it.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
For certain arches (AARCH64, x86, generic) get_pixels and
get_pixels_unaligned always coincide for 8 bit input.
In these cases it is possible to avoid checks for unaligned
input in asvenc, dvenc.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Before this commit, the input in get_pixels and get_pixels_unaligned
has been treated inconsistenly:
- The generic code treated 9, 10, 12 and 14 bits as 16bit input
(these bits correspond to what FFmpeg's dsputils supported),
everything with <= 8 bits as 8 bit and everything else as 8 bit
when used via AVDCT (which exposes these functions and purports
to support up to 14 bits).
- AARCH64, ARM, PPC and RISC-V, x86 ignore this AVDCT special case.
- RISC-V also ignored the restriction to 9, 10, 12 and 14 for its
16bit check and treated everything > 8 bits as 16bit.
- The mmi MIPS code treats everything as 8 bit when used via
AVDCT (this is certainly broken); otherwise it checks for <= 8 bits.
The msa MIPS code behaves like the generic code.
This commit changes this to treat 9..16 bits as 16 bit input,
everything else as 8 bit (the former because it makes sense,
the latter to preserve the behaviour for external users*).
*: The only internal user of AVDCT (the spp filter) always
uses 8, 9 or 10 bits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Fixes segfaults on systems where PixblockDSPContext.get_pixels
really requires to be properly aligned (e.g. ARMv7).
Before this commit input created by
-filter_complex nullsrc=s=740x576:r=25,format=yuv420p,crop=w=720:x=2
led to crashes.
(The unaligned strides are in violation of the AVFrame.linesize
documentation, unaligned pointers itself do not seem to be
prohibited for encoders.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
