With certain types of input and the filter chain getting re-initialized
or re-configured, multiple nullptr AVSubtitles can get pushed into
sub2video_update() in a row from sub2video_heartbeat.

This causes end_pts, and on the next round pts, to become INT64_MAX,
latter of which signals EOF in framesync, leading to complete loss of
subtitles from that point on.

Thus, check that the sub2video.end_pts is smaller than INT64_MAX
in a similar fashion to sub2video_flush before sending out the
nullptr AVSubtitle. This keeps premature EOFs from happening in
framesync and the subtitle overlay is kept past the filter chain
re-initializations/configurations.

On modern x86 systems its around 2x faster. For systems without
FPUs it'll be slower, but our policy is to prefer floating point
implementations and to let users decide what's best (or just not
compile them on systems without FPUs).

Signed-off-by: Rostislav Pehlivanov <[email protected]>

and use it if fade == 0.

… the loop

…_list()

example:

AVPixelFormat pixFmts[] = { AV_PIX_FMT_RGB24, AV_PIX_FMT_RGBA };
int loss = 0;
AVPixelFormat best = avcodec_find_best_pix_fmt_of_list(pixFmts, AV_PIX_FMT_BGRA, 1, &loss);

best is AV_PIX_FMT_RGB24. But AV_PIX_FMT_RGBA is better.

Signed-off-by: Michael Niedermayer <[email protected]>

Signed-off-by: Paul B Mahol <[email protected]>

Fixes #6563.

Signed-off-by: Paul B Mahol <[email protected]>

Fixes #4889.

Signed-off-by: Paul B Mahol <[email protected]>

Signed-off-by: Paul B Mahol <[email protected]>

Signed-off-by: Paul B Mahol <[email protected]>

Signed-off-by: Josh de Kock <[email protected]>

Signed-off-by: Josh de Kock <[email protected]>

fix commit 2238190 break the "ffmpeg -muxers/demuxers".

Signed-off-by: Jun Zhao <[email protected]>
Signed-off-by: Josh de Kock <[email protected]>

Support for it was apparently never in the codebase, and the enum
values were recently removed from the public headers [1]

Fixes build with latest libaom build.

[1] https://aomedia.googlesource.com/aom/+/2e3cd5c5c30fa02134681cda900c32486807af3f

Signed-off-by: James Almer <[email protected]>

And fix the entry in doc/APIchanges

Signed-off-by: James Almer <[email protected]>

This reverts commit 909e00a.

There is no need to use the old API anymore as the new API now
behaves in the same way (treating devices as formats when loaded).

Signed-off-by: Josh de Kock <[email protected]>

Signed-off-by: Josh de Kock <[email protected]>

avdevice_register_all() is still required to register devices into
lavf (this is required due to lavd being somewhat of a hack).

Signed-off-by: Josh de Kock <[email protected]>

Signed-off-by: Josh de Kock <[email protected]>

It works as a drop in replacement for the deprecated av_dup_packet(),
to ensure a packet is reference counted.

Reviewed-by: Michael Niedermayer <[email protected]>
Signed-off-by: James Almer <[email protected]>

Signed-off-by: James Almer <[email protected]>

…ef counted

Simplifies code.

Signed-off-by: James Almer <[email protected]>

…s are ref counted

Partially reverts commit e91f0c4, simplifying code.

Signed-off-by: James Almer <[email protected]>

…counted

Simplifies code.

Signed-off-by: James Almer <[email protected]>

…ef counted

Simplifies code, while also fixing a potential leak of side data in pkt.

Signed-off-by: James Almer <[email protected]>

Signed-off-by: Paul B Mahol <[email protected]>

This is a profile supporting > 8-bit video and has a higher quality DCT

Signed-off-by: Paul B Mahol <[email protected]>

Fate output changes and under close inspection looks more correct.
Fixes #6187.

Signed-off-by: Paul B Mahol <[email protected]>

Fixes ticket #6434.

Reviewed-by: Nicolas George <[email protected]>
Signed-off-by: Marton Balint <[email protected]>

Also do not set empty metadata.

Signed-off-by: Paul B Mahol <[email protected]>

Support for this needs testing, so remove for now.

Signed-off-by: James Almer <[email protected]>

Support for this needs testing, so remove for now.

Signed-off-by: James Almer <[email protected]>

Change the seek2any location in avformat_options to make code more
readable.

Signed-off-by: Jun Zhao <[email protected]>
Signed-off-by: Michael Niedermayer <[email protected]>

Fixes: Timeout
Fixes: 6304/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_CSCD_fuzzer-5754772461191168

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <[email protected]>

…ing_fixed()

I was not able to reproduce this, this fix is based on just the fuzzer log.
Fixes: 4959/clusterfuzz-testcase-minimized-6035350934781952

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <[email protected]>

Allows to manage the AV_LOG_PRINT_LEVEL flag as a prefix to the loglevel
option value, similar to the existing AV_LOG_SKIP_REPEATE flag. Adds
support for setting flags relative to the existing value by using a +/-
prefix.

Previous version reviewed-by: Michael Niedermayer <[email protected]>
Signed-off-by: Tobias Rapp <[email protected]>

regression since 354b26a

Signed-off-by: Paul B Mahol <[email protected]>

Signed-off-by: Paul B Mahol <[email protected]>

This is the only code I found within FFmpeg that still inherently
requires AV_PIX_FMT_FLAG_PSEUDOPAL. It's easily changed not to require
it.

Preparation for the next patch.

PSEUDOPAL pixel formats are not paletted, but carried a palette with the
intention of allowing code to treat unpaletted formats as paletted. The
palette simply mapped the byte values to the resulting RGB values,
making it some sort of LUT for RGB conversion.

It was used for 1 byte formats only: RGB4_BYTE, BGR4_BYTE, RGB8, BGR8,
GRAY8. The first 4 are awfully obscure, used only by some ancient bitmap
formats. The last one, GRAY8, is more common, but its treatment is
grossly incorrect. It considers full range GRAY8 only, so GRAY8 coming
from typical Y video planes was not mapped to the correct RGB values.
This cannot be fixed, because AVFrame.color_range can be freely changed
at runtime, and there is nothing to ensure the pseudo palette is
updated.

Also, nothing actually used the PSEUDOPAL palette data, except xwdenc
(trivially changed in the previous commit). All other code had to treat
it as a special case, just to ignore or to propagate palette data.

In conclusion, this was just a very strange old mechnaism that has no
real justification to exist anymore (although it may have been nice and
useful in the past). Now it's an artifact that makes the API harder to
use: API users who allocate their own pixel data have to be aware that
they need to allocate the palette, or FFmpeg will crash on them in
_some_ situations. On top of this, there was no API to allocate the
pseuo palette outside of av_frame_get_buffer().

This patch not only deprecates AV_PIX_FMT_FLAG_PSEUDOPAL, but also makes
the pseudo palette optional. Nothing accesses it anymore, though if it's
set, it's propagated. It's still allocated and initialized for
compatibility with API users that rely on this feature. But new API
users do not need to allocate it. This was an explicit goal of this
patch.

Most changes replace AV_PIX_FMT_FLAG_PSEUDOPAL with FF_PSEUDOPAL. I
first tried #ifdefing all code, but it was a mess. The FF_PSEUDOPAL
macro reduces the mess, and still allows defining FF_API_PSEUDOPAL to 0.

Passes FATE with FF_API_PSEUDOPAL enabled and disabled. In addition,
FATE passes with FF_API_PSEUDOPAL set to 1, but with allocation
functions manually changed to not allocating a palette.

…le of 16

Also update fate reference. Fixes ticket #6675.

Signed-off-by: James Almer <[email protected]>

All such files are signaled as I42016, as there's no monochrome value
in aom_img_fmt_t.

Signed-off-by: James Almer <[email protected]>

Fixes make fate-source

Signed-off-by: James Almer <[email protected]>

Fixes make fate-source

Signed-off-by: James Almer <[email protected]>

Signed-off-by: Jun Zhao <[email protected]>
Reviewed-by: Steven Liu <[email protected]>

Based on a patch by Luca Barbato.

Signed-off-by: James Almer <[email protected]>

Don't constantly overwrite the list's head pointer.

Signed-off-by: James Almer <[email protected]>

It's more robust and efficient.

Signed-off-by: James Almer <[email protected]>

Simplifies code.

Signed-off-by: James Almer <[email protected]>

Simplifies code.

Reviewed-by: Michael Niedermayer <[email protected]>
Signed-off-by: James Almer <[email protected]>

Signed-off-by: James Almer <[email protected]>

Fixes memleaks.

Signed-off-by: James Almer <[email protected]>

Signed-off-by: James Almer <[email protected]>

Fixes memleaks.

Signed-off-by: James Almer <[email protected]>

Should fix double free related crashes.

Signed-off-by: James Almer <[email protected]>

Fixes memleaks.

Signed-off-by: James Almer <[email protected]>

This is used to signal that image should be stored in metadata
as cover image.

Signed-off-by: Timo Teräs <[email protected]>
Signed-off-by: Michael Niedermayer <[email protected]>

This is shared test and this simplifies code a bit. Follow up
commit will have additional tests for this function.

Signed-off-by: Timo Teräs <[email protected]>
Signed-off-by: Michael Niedermayer <[email protected]>

Support for it was apparently never in the codebase, and the enum
values were recently removed from the public headers [1]

Fixes build with latest libaom build.

[1] https://aomedia.googlesource.com/aom/+/3f29cc20e3a4c348cb41a797c68de856ddb84e12

Signed-off-by: James Almer <[email protected]>

The pseudo palette allocation is optional now. But if it's still
allocated (like the internal get_buffer2 implementation does, for
compatibility), it shouldn't print a warning.

Found by Hendrik Schreiber.

Signed-off-by: Lou Logan <[email protected]>

Signed-off-by: Jacob Trimble <[email protected]>
Signed-off-by: Michael Niedermayer <[email protected]>

…_type

Signed-off-by: Valery Kot <[email protected]>

…plication between 8bit and 16 bit version

grainextract
grainmerge
average
extremity
negation

grainextract
grainmerge
average
extremity
negation

… step

also remove unused arg in estimate_alpha_plane

Improve quality of Prores 4444 encoding

Signed-off-by: Josh de Kock <[email protected]>

Signed-off-by: Josh de Kock <[email protected]>

Fixes ticket #7125.

Info about default value of bind_address option and its abbreviated
version (b). Example modified to have named instanced filter and to show
its use.

Signed-off-by: Bela Bodecs <[email protected]>
Signed-off-by: Lou Logan <[email protected]>

This seems to confuse Github users into thinking that we may accept pull
requests. We do not accept pull requests.

Sending patches to the ffmpeg-devel mailing list is our preferred method
for users to contribute code.

Signed-off-by: Lou Logan <[email protected]>

… sum2_float during dithering.

Removed +len1 in call to s->mix_2_1_f() as I found no logical explanation for it. After removal, problem was gone.

Signed-off-by: Hendrik Schreiber <[email protected]>
Signed-off-by: Michael Niedermayer <[email protected]>

Data in EbmlBin objects is never changed after being read from the
input file (save for two specific cases with encoded CodePrivate), so
using AVBufferRef we can prevent unnecessary copy of data by instead
creating new references to said constant data.

Signed-off-by: James Almer <[email protected]>

Simplifies code in matroska_parse_frame(). This is in preparation for
the following patch.

Reviewed-by: Michael Niedermayer <[email protected]>
Signed-off-by: James Almer <[email protected]>

…g packets

Newly allocated data buffers (wavpack, prores, compressed buffers)
are padded to meet the requirements of AVPacket.

About 10x speed up in matroska_parse_frame().

Reviewed-by: Michael Niedermayer <[email protected]>
Signed-off-by: James Almer <[email protected]>

Found-by: Kieran
Signed-off-by: Michael Niedermayer <[email protected]>

Fixes visual corruption on seeking

Fixes: downloadTest_clip_24M.rmvb

Signed-off-by: Michael Niedermayer <[email protected]>

Fixes: runtime error: signed integer overflow: 2052929346 + 204817098 cannot be represented in type 'int'

This was missed in b1bef75
Fixes: 5275/clusterfuzz-testcase-minimized-5367635958038528

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <[email protected]>

Currently pict_type and key_frame are unset.
Add an extra param to fetch the picture type from qsv decoder

The judgement “key frame is equal to IDR frame” only suitable for H264.
For HEVC, all IRAP frames are key frames, and other codecs have no IDR
frame.

Signed-off-by: ChaoX A Liu <[email protected]>
Signed-off-by: Zhong Li <[email protected]>
Signed-off-by: Maxym Dmytrychenko <[email protected]>

Currently vpp pipeline is always created, even for the unnecessary
cases such as setting the option "vpp_qsv=w=1280:h=720" for an input
with native resolution 1280x720. Thus introduces unnecessary performance
dropping, so bypass vpp if not needed.

Signed-off-by: Zhong Li <[email protected]>
Signed-off-by: Maxym Dmytrychenko <[email protected]>

Starting from API 1.25 helps to improve performance of the simultaneous
encode, 1:N scenario, like:

./avconv  -y -hwaccel qsv -c:v h264_qsv -r 30000/1001 -i
~/bbb_sunflower_1080p_60fps_normal.mp4  -vframes 600 -an \
    -filter_complex "split=2[s1][s2]; [s1]scale_qsv=1280:720[o1];
[s2]scale_qsv=960:540[o2]" \
    -map [o1] -c:v h264_qsv -b:v 3200k -minrate 3200k -maxrate 3200k -f
rawvideo /tmp/3200a.264 \
    -map [o2] -c:v h264_qsv -b:v 1750k -minrate 1750k -maxrate 1750k -f
rawvideo /tmp/1750a.264

Signed-off-by: Maxym Dmytrychenko <[email protected]>

Set the flag MFX_BITSTREAM_COMPLETE_FRAME when it is a progressive picture.
This can fix vc1 decoding segment fault issues because can't set the start
code correctly.

See: ./avconv -hwaccel qsv -c:v vc1_qsv -i /fate-suite/vc1/SA00040.vc1
-vf "hwdownload, format=nv12" -f rawvideo /dev/null

v2: fix some h264 interlaced clips regression
a. field_order of some h264 interlaced video (e.g: cama3_vtc_b.avc) is marked as AV_FIELD_UNKNOWN
   in h264_parser.c. This is not a completed frames.
   So only set the MFX_BITSTREAM_COMPLETE_FRAME when it is progressive.
b. some clips have both progressive and interlaced frames (e.g.CAPAMA3_Sand_F.264),
   the parsed field_order maybe changed druing the decoding progress.

This patch has been verified for other codecs(mpeg2/hevc/vp8).

Signed-off-by: Zhong Li <[email protected]>
Signed-off-by: Maxym Dmytrychenko <[email protected]>

Limits based on 3GPP TS 26.245 V14.0.0
Fixes: Timeout
Fixes: 6377/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MOVTEXT_fuzzer-5175929115508736

Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Reviewed-by: Philip Langdale <[email protected]>
Signed-off-by: Michael Niedermayer <[email protected]>

Signed-off-by: Meng Wang <[email protected]>
Reviewed-by: Shengbin Meng <[email protected]>
Signed-off-by: Michael Niedermayer <[email protected]>

Where tiny is less than the default CTU size.

Signed-off-by: Vittorio Giovara <[email protected]>

Shift fixed point numbers to be actual decimal numbers.

This removes the XP compatibility code, and switches entirely to SRW
locks, which are available starting at Windows Vista.

This removes CRITICAL_SECTION use, which allows us to add
PTHREAD_MUTEX_INITIALIZER, which will be useful later.

Windows XP is hereby not a supported build target anymore.

Signed-off-by: Diego Biurrun <[email protected]>

Fixes two warnings: ISO C90 forbids mixed declarations and code

Signed-off-by: Michael Niedermayer <[email protected]>