1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-02 03:06:28 +02:00
Commit Graph

1451 Commits

Author SHA1 Message Date
Anton Khirnov
3bd7c57125 fftools/ffmpeg_filter: implement filtergraph chaining
This allows one complex filtergraph's output to be sent as input to
another one, which is useful in certain situations (one is described in
the docs).

Chaining filtergraphs was already effectively possible by using a
wrapped_avframe encoder connected to a loopback decoder, but it is ugly,
non-obvious and inefficient.
2024-04-09 10:34:18 +02:00
Anton Khirnov
255ae03601 fftools/ffmpeg_sched: allow filtergraphs to send to filtergraphs
Will be useful for filtergraph chaining that will be added in following
commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
3d01996b24 fftools/ffmpeg_filter: change processing order in fg_finalise_bindings()
First bind all inputs in all filtergraphs, only then check that all
outputs are bound.

Needed by the following commit.
2024-04-09 10:34:18 +02:00
Anton Khirnov
243a51490a fftools/ffmpeg_filter: only store complex filtergraphs in global array
Store simple filtergraphs in the stream they feed. Keeping the two
separate will be useful in following commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
d74cbcb963 fftools/ffmpeg_filter: drop OutputFilter.ost
All remaining code accessing it only needs to know whether this
filtergraph output has been bound or not.
2024-04-09 10:34:18 +02:00
Anton Khirnov
f2c919252d fftools/ffmpeg_filter: accept encoder thread count through OutputFilterOptions
Stop digging through encoder options manually.

Will allow decoupling filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
bfeb751171 fftools/ffmpeg_mux: drop OutputFile.format
It is no longer used outside of the muxing code (where we can access the
muxer directly).
2024-04-09 10:34:18 +02:00
Anton Khirnov
5b0589c8c3 fftools/ffmpeg_mux: drop OutputFile.shortest
It is no longer needed outside of of_open() and its children.
2024-04-09 10:34:18 +02:00
Anton Khirnov
a4c940c86a fftools/ffmpeg_filter: move most of -apad logic to the muxer
The decision whether -apad actually does anything is made based on muxer
properties, and so more properly belongs there. Filtering code only
receives the result.
2024-04-09 10:34:18 +02:00
Anton Khirnov
a2892dbe06 fftools/ffmpeg_filter: pass trim parameters through OutputFilterOptions
Do not read them from OutputStream directly.

Will allow decoupling filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
83304f7c1f fftools/ffmpeg_filter: pass autoscale through OutputFilterOptions
Do not read it from OutputStream directly.

Will allow decoupling filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
b8e6802023 fftools/ffmpeg_filter: pass sws/swr opts through OutputFilterOptions
Do not read them from OutputStream directly.

Will allow decoupling filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
23c23077fc fftools/ffmpeg_filter: drop an unnecessary use of OutputStream
OutputFilter.type contains the same information.
2024-04-09 10:34:18 +02:00
Anton Khirnov
fc6354c39c fftools/ffmpeg_filter: add an AVClass to OutputFilter
Use it for logging where appropriate, avoid logging to OutputStream as
we do not own it.

This is a step towards decoupling filtering from encoding.
2024-04-09 10:34:18 +02:00
Anton Khirnov
3c3e04c8a3 fftools/ffmpeg_filter: simplify retrieving filter type 2024-04-09 10:34:18 +02:00
Anton Khirnov
114cbaa316 fftools/ffmpeg_filter: drop a redundant check
fg_finalise_bindings() already checks that all filtergraph outputs are
connected.
2024-04-09 10:34:18 +02:00
Anton Khirnov
da80e0b077 fftools/ffmpeg_filter: accept a caller-provided output name
Do not construct it from OutputStream manually.

Will allow decoupling filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
82c7c21b18 fftools/ffmpeg: drop OutputStream.is_cfr
It is used in a single place in the filtering code, so it is better to
inline it there.
2024-04-09 10:34:18 +02:00
Anton Khirnov
509afedaaf fftools/ffmpeg_filter: pass vsync method through OutputFilterOptions
Do not read it from OutputStream directly.

Will allow decoupling filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
bc206ed1b3 fftools/ffmpeg_filter: stop accessing encoder AVCodecContext
Pass all the necessary value through OutputFilterOptions.

Will allow decoupling filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
17702c5f7b fftools/ffmpeg_filter: move the MJPEG format selection hack to muxer setup
That, if anywhere, is a more appropriate place for it.
2024-04-09 10:34:18 +02:00
Anton Khirnov
9d5bf2d69e fftools/ffmpeg_filter: pass enc_timebase through OutputFilterOptions
Reduces the need to access OutputStream, which will allow decoupling
filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
606c71bb11 fftools/ffmpeg: warn about ignored -enc_time_base for subtitles earlier
Can do it as soon as that option is parsed, no need to postpone it until
opening the encoder.
2024-04-09 10:34:18 +02:00
Anton Khirnov
e903c31fd1 fftools/ffmpeg_filter: pass keep_pix_fmt through OutputFilterOptions
Reduces the need to access OutputStream, which will allow decoupling
filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
b3864e7a08 fftools/ffmpeg_filter: pass ts offset through OutputFilterOptions
Reduces the need to access OutputFile, which will allow decoupling
filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
651c79da36 fftools/ffmpeg_filter: check that filter type matches output stream type
For simple filtergraphs. For complex filtergraphs they always match.
2024-04-09 10:34:18 +02:00
Anton Khirnov
8e35e33d42 fftools/ffmpeg_filter: stop accessing AVCodecContext.codec
Instead pass the encoder through a newly-added output options struct,
analogous to previously added input options.

Will allow decoupling filtering from encoding in future commits.
2024-04-09 10:34:18 +02:00
Anton Khirnov
8c33085313 fftools/ffmpeg_filter: do not pass OutputStream to set_channel_layout()
It only needs a list of allowed layouts and the requested layout.
2024-04-09 10:34:18 +02:00
Andreas Rheinhardt
790f793844 avutil/common: Don't auto-include mem.h
There are lots of files that don't need it: The number of object
files that actually need it went down from 2011 to 884 here.

Keep it for external users in order to not cause breakages.

Also improve the other headers a bit while just at it.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2024-03-31 00:08:43 +01:00
Anton Khirnov
24b9f29ff2 fftools/ffmpeg_sched: make sure to always run task cleanup
Even in cases where sch_start() failed. This ensures all links are
properly closed and no tasks are left hanging.

Fixes #10916.
2024-03-29 08:51:12 +01:00
Anton Khirnov
af81788f30 fftools/ffmpeg_sched: move sch_stop() to the bottom of the file
Will allow avoiding forward declarations in following commits.
2024-03-29 08:51:12 +01:00
Andreas Rheinhardt
5d71f97e0e all: Don't use ATOMIC_VAR_INIT
C11 required to use ATOMIC_VAR_INIT to statically initialize
atomic objects with static storage duration. Yet this macro
was unsuitable for initializing structures [1] and was actually
unneeded for all known implementations (this includes our
compatibility fallback implementations which simply wrap the value
in parentheses: #define ATOMIC_VAR_INIT(value) (value)).
Therefore C17 deprecated the macro and C23 actually removed it [2].

Since commit 5ff0eb34d2 we default
to C17 if the compiler supports it; Clang warns about ATOMIC_VAR_INIT
in this mode. Given that no implementation ever needed this macro,
this commit stops using it to avoid this warning.

[1]: https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2396.htm#dr_485
[2]: https://en.cppreference.com/w/c/atomic/ATOMIC_VAR_INIT

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2024-03-28 09:12:48 +01:00
Anton Khirnov
a69cedb6a6 fftools/ffmpeg_demux: make InputStream.autorotate private
It is no longer accessed outside of ffmpeg_demux.
2024-03-28 08:40:11 +01:00
Anton Khirnov
c9a90c052a fftools/ffmpeg_enc: stop copying demuxer side data to the muxer
All side data should be propagated through the trancoding pipeline.
2024-03-28 08:40:11 +01:00
Anton Khirnov
106131bb10 fftools/ffmpeg_filter: remove display matrix if we have applied it 2024-03-28 08:40:01 +01:00
Anton Khirnov
b1aaa1f585 fftools/ffmpeg_dec: apply cropping manually
lavfi does not require aligned buffers, so we can safely apply top/left
cropping by any amount, without passing any special flags to lavc.
Longer term, an even better solution would probably be auto-inserting
the crop filter (or its hwaccel versions) as needed.

Multiple FATE tests no longer need -flags unaligned.
2024-03-28 08:40:01 +01:00
Anton Khirnov
0edbd00ccf fftools/ffmpeg_{demux,dec}: pass -bitexact through DecoderFlags
Avoids abusing AV_DICT_MULTIKEY and relying on undocumented AVDictionary
ordering behaviour.
2024-03-28 08:40:01 +01:00
Anton Khirnov
372c78dd42 fftools/ffmpeg_dec: apply decoder options manually
Do not pass an options dictionary to avcodec_open2().

This should be equivalent to current behaviour, but will allow
overriding caller-supplied options in a cleaner and more robust manner.

We can now set the COPY_OPAQUE flag directly rather going through
dec_opts.
2024-03-28 08:40:01 +01:00
Anton Khirnov
2d06a7570e fftools/cmdutils: do not use a random codec's private options
There is only a single caller of filter_codec_opts() that passes
a NULL codec to it, which is streamcopy in ffmpeg CLI. In that case we
only want generic AVCodecContext options, not private options of any
specific encoder.
2024-03-28 08:40:01 +01:00
Anton Khirnov
da67816128 fftools/ffmpeg_demux: only call filter_codec_opts() when we have a decoder
It is pointless otherwise, as decoder options will not be used.
2024-03-28 08:40:01 +01:00
Andreas Rheinhardt
432e287e27 fftools/ffmpeg_sched: Explicitly return 0 on sch_enc_send() success
Do not return the return value of the last enc_send_to_dst()
call, as this would treat the last call differently from the
earlier calls; furthermore, sch_enc_send() explicitly documents
to always return 0 on success.

Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2024-03-28 03:06:13 +01:00
Anton Khirnov
8fc1e1358b fftools/ffmpeg_enc: simplify error handling for decoded_side_data setup
There is no need to free the already-added items, they will be freed
alongside the codec context. There is also little point in an error
message, as the only reason this can fail is malloc failure.
2024-03-27 11:36:21 +01:00
Anton Khirnov
6f2cb0923c fftools/ffmpeg_enc: move decoded_side_data setup out of video-only block
Nothing about this code is video-specific.
2024-03-27 11:36:20 +01:00
Anton Khirnov
fabf148578 fftools/ffmpeg_enc: only promote first frame side data to global when meaningful
Skip those side data types that do not make sense as global side data.
2024-03-27 11:35:27 +01:00
Andreas Rheinhardt
ced5c5fdb8 fftools/ffmpeg_mux_init: Fix double-free on error
MATCH_PER_STREAM_OPT iterates over all options of a given
OptionDef and tests whether they apply to the current stream;
if so, they are set to ost->apad, otherwise, the code errors
out. If no error happens, ost->apad is av_strdup'ed in order
to take ownership of this pointer.

But this means that setting it originally was premature,
as it leads to double-frees when an error happens lateron.
This can simply be reproduced with
ffmpeg -filter_complex anullsrc  -apad bar -apad:n baz -f null -
This is a regression since 83ace80bfd.

Fix this by using a temporary variable instead of directly
setting ost->apad. Also only strdup the string if it actually
is != NULL.

Reviewed-by: Marth64 <marth64@proxyid.net>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2024-03-26 06:48:35 +01:00
James Almer
0963ef4996 fftools/ffmpeg_filter: remove prototype for non existent function
Signed-off-by: James Almer <jamrial@gmail.com>
2024-03-25 23:23:27 -03:00
James Almer
767e7d3d2b fftools/ffmpeg_filter: remove unused struct from InputFilterPriv
It's already in InputFilterOptions.

Signed-off-by: James Almer <jamrial@gmail.com>
2024-03-25 23:23:27 -03:00
Niklas Haas
6963033590 ffprobe: adapt to new AVFilmGrainParams
Follow the establish convention of printing the bit depth metadata
per-component.
2024-03-23 18:54:36 +01:00
Andreas Rheinhardt
c77164390b fftools/ffmpeg_enc: Don't call frame_data twice
Reviewed-by: Jan Ekström <jeebjp@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2024-03-22 22:11:04 +01:00
Jan Ekström
8f4b173029 ffmpeg: pass first video AVFrame's side data to encoder
This enables further configuration of output based on the results
of input decoding and filtering in a similar manner as the color
information.
2024-03-20 19:15:05 +02:00