The rationale is that coded_frame was only used to communicate key_frame,
pict_type and quality to the caller, as well as a few other random fields,
in a non predictable, let alone consistent way.
There was agreement that there was no use case for coded_frame, as it is
a full-sized AVFrame container used for just 2-3 int-sized properties,
which shouldn't even belong into the AVCodecContext in the first place.
The appropriate AVPacket flag can be used instead of key_frame, while
quality is exported with the new AVPacketSideData quality factor.
There is no replacement for the other fields as they were unreliable,
mishandled or just not used at all.
This is necessary to preserve the quality information currently exported
with coded_frame. Add the new side data to every encoder that needs it,
and use it in avconv.
Gather all coded_frame allocations and free functions to a single place
Allocating coded_frame is what most encoders do anyway, so it makes
sense to always allocate and free it in a single place. Moreover a lot
of encoders freed the frame with av_freep() instead of the correct API
av_frame_free().
This bring uniformity to encoder behaviour and prevents applications
from erroneusly accessing this field when not allocated. Additionally
this helps isolating encoders that export information with coded_frame,
and heavily simplifies its deprecation.
This change (and the following ones of the same kind) is mainly to
simplify wrapping this section with an #if FF_API block later on.
No functional changes are applied, the fields of the context coded_frame
fields are directly initialized, instead of keeping a reference to the
coded_frame itself.
Anton Khirnov [Thu, 16 Jul 2015 17:26:36 +0000 (19:26 +0200)]
af_resample: do not touch the timestamps if we are not resampling
This filter currently assumes that the input audio is continuous and
does some timestamps manipulation based on this assumption.
This is unnecessary if we are only converting the channel layout or the
sample format, without resampling. In such a case, just leave the
timestamps as they are.
Anton Khirnov [Wed, 15 Jul 2015 17:49:24 +0000 (19:49 +0200)]
avconv: split creating and (re-)configuring complex filtergraphs
The current code is less than straightforward due to the fact that
output streams can be created based on filtergraph definitions. This
change should make the code simpler and more readable. It will also be
useful in the future commits.
Anton Khirnov [Wed, 15 Jul 2015 15:17:54 +0000 (17:17 +0200)]
avconv: create the complex filtergraphs earlier
Since global options are processed before all the other options now, we
do not have to try creating the complex filtergraphs several times
anymore, it is enough to do it once after the input files are opened.
Anton Khirnov [Sun, 28 Jun 2015 05:41:22 +0000 (07:41 +0200)]
avconv: drop update_sample_fmt()
There is only one decoder left that supports this (libopus, which is not
used by default since we have a native one) and this code goes against
the avconv design, since it propagates information back from the encoder
to decoder.
Henrik Gramner [Wed, 15 Jul 2015 22:10:27 +0000 (00:10 +0200)]
x86: bswapdsp: Don't treat 32-bit integers as 64-bit
The upper halves are not guaranteed to be zero in x86-64.
Also use `test` instead of `and` when the result isn't used for anything other
than as a branch condition, this allows some register moves to be eliminated.
Dave Yeo [Fri, 17 Jul 2015 05:21:11 +0000 (22:21 -0700)]
OS/2: Cleanup slib_create_def_cmd
Export symbols by name rather then ordinal.
Remove PROTMODE directive as it does not make sense for 32 bit library.
Also silences a warning from some linkers.
Signed-off-by: Dave Yeo <dave.r.yeo@gmail.com> Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Anton Khirnov [Fri, 10 Jul 2015 07:31:24 +0000 (09:31 +0200)]
bytestream2: set the reader to the end when reading more than available
This prevents possible infinite loops with the calling code along the
lines of while (bytestream2_get_bytes_left()) { ... }, where the reader
does not advance.
Henrik Gramner [Sat, 11 Jul 2015 18:32:11 +0000 (20:32 +0200)]
Checkasm: assembly testing and benchmarking tool
It provides the following features:
* verify correctness by comparing output to the C version.
* detect failure to save and restore clobbered callee-saved registers.
* detect 32-bit parameters being used as if they were 64-bit in x86-64
(the upper halves are not guaranteed to be zero - but in practice
they very often are, which makes those bugs hard to spot otherwise).
* easy benchmarking.
Compile by running 'make checkasm'.
Execute by running 'tests/checkasm/checkasm'.
Optional arguments are '--bench' to run benchmarks for all functions,
'--bench=<pattern>' to run benchmarks for all functions that starts with
<pattern>, and '<integer>' to seed the PRNG for reproducible results.
Contains unit tests for most h264pred functions to get started, more tests
can be added afterwards using those as a reference.
Loosely based on code from x264. Currently only supports x86 and x86-64,
but additional architectures shouldn't be too much of an obstacle to add.
Note that functions with floating point parameters or floating point
return values are not supported. Some compiler-specific features or
preprocessor hacks would likely be required to add support for that.
Signed-off-by: Janne Grunau <janne-libav@jannau.net>
Henrik Gramner [Wed, 8 Jul 2015 20:33:53 +0000 (22:33 +0200)]
x86: Serialize rdtsc in read_time()
Improves the accuracy of measurements, especially in short sections.
To quote the Intel 64 and IA-32 Architectures Software Developer's Manual:
"The RDTSC instruction is not a serializing instruction. It does not necessarily
wait until all previous instructions have been executed before reading the counter.
Similarly, subsequent instructions may begin execution before the read operation
is performed. If software requires RDTSC to be executed only after all previous
instructions have completed locally, it can either use RDTSCP (if the processor
supports that instruction) or execute the sequence LFENCE;RDTSC."
SSE2 is a requirement for lfence so only use it on SSE2-capable systems.
Prefer lfence;rdtsc over rdtscp since rdtscp is supported on fewer systems.
avconv: vda: Unlock the pixel buffer once it is accessed
Avoid possible issues with memmapped hardware buffers in
case VDA is not doing a conversion on behalf of the user
and make the code more proper as working example.
Stian Selnes [Tue, 2 Dec 2014 17:55:13 +0000 (18:55 +0100)]
h261: Signal freeze picture release for intra frames
Freeze picture release should be set to 1 when we're responding to a
fast update request. For simplicity we set it for all intra frames,
including those that starts a GOP.
Fixes issue where Tandberg MXP1700 does not recover from packet loss
state since it's waiting for the freeze picture relase indication.
Stian Selnes [Fri, 12 Dec 2014 15:08:19 +0000 (16:08 +0100)]
h261: Set 'still image mode off' in picture header
Ref H.261 recommendation section 4.2.1.3, setting the still image flag
to 1 disables still image mode. Some decoders require this in order to
decode the bitstream as normal video.
Fixes H.261 calls to Cisco E20.
Also, reserved (aka spare) bits should be set to 1 unless specified
otherwise.
projects / ffmpeg / log