Given that the AVCodec.next pointer has now been removed, most of the
AVCodecs are not modified at all any more and can therefore be made
const (as this patch does); the only exceptions are the very few codecs
for external libraries that have a init_static_data callback.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>
Explicitly identify decoder/encoder wrappers with a common name. This
saves API users from guessing by the name suffix. For example, they
don't have to guess that "h264_qsv" is the h264 QSV implementation, and
instead they can just check the AVCodec .codec and .wrapper_name fields.
Explicitly mark AVCodec entries that are hardware decoders or most
likely hardware decoders with new AV_CODEC_CAPs. The purpose is allowing
API users listing hardware decoders in a more generic way. The proposed
AVCodecHWConfig does not provide this information fully, because it's
concerned with decoder configuration, not information about the fact
whether the hardware is used or not.
AV_CODEC_CAP_HYBRID exists specifically for QSV, which can have software
implementations in case the hardware is not capable.
Based on a patch by Philip Langdale <philipl@overt.org>.
Merges Libav commit 47687a2f8a.
I tried doing this before, but it resulted in weird behaviour with
certain samples. I want to say I think I've got it sorted out now,
and the new autobsf stuff makes it trivial to turn on.
The native support for packed bframes is buggy and I think buggy
in ways beyond what I already try to account for, so this should be
a net improvements.
Previously, the pts value was initialised to AV_NOPTS_VALUE and so
it was not necessary to always set it. Now, with the new-new decode
API, this is no longer true. I'm not sure why I avoided setting the
pts when the decoder value was also AV_NOPTS_VALUE - it clearly
wouldn't have changed anything previously, but here we are.
Failing to do this, means the frame pts will be some random uninitalised
value.
The new new decode API requires the decoder to ask for the next input
packet, and it cannot just return EAGAIN if that packet cannot be
processed yet. This means we must finally confront how we get this
decoder to block when the input buffer is full and no output frames
are ready yet.
In the end, that isn't too hard to achieve - the main trick seems to
be that you have to aggressively poll the hardware - it doesn't seem
to make any forward progress if you sleep.
Signed-off-by: James Almer <jamrial@gmail.com>
The old code had to retain a partial frame across two calls in
the case of separate interlaced fields. Now, we know that we'll
get both fields within the same receive_frame call, and so we
don't need to manage the frame as private state any more.
It's not possible to return EAGAIN when we've passed input EOF and are
in draining mode. If do return EAGAIN, we're saying there's no way to
get any more output - which isn't true in many cases.
So let's handled these cases in an internal loop as best we can.
It seems that without all the other 1:1 heuristics, we don't have
a fundamental problem trusting the interlaced flag on output
pictures. That's a relief.
I'm not sure why, but the mpeg4_unpack_bframes bsf is not
interacting well with seeking. Looking at the code, it should be
ok, with possibly one warning shown, but I see it getting stuck
for an extended period of time after a seek where a packed frame
is cached to be shown later.
So, I gave up on that and went back to making the old hardware
based path work. Turns out that it wasn't broken except that some
samples have a 6 byte drop packet which I wasn't accounting for.
Now it works again and seeks are good.
The new decode API allows for m:n decode patterns, which is what
you need to use this hardware in a sane way. There are so many
situations where 1:1 doesn't happen naturally that it's a miracle
I got it working as well as I did.
With this change, we can throw all of the crazy heuristics and
sleeps(!) out, and things work correctly.
Why on earth the hardware returns garbage for the first sample of
a decoded picture is anyone's guess. The simplest reasonable way
to patch it up is to copy the first sample of the second line. This
should result in the correct chroma values (because the data was
original 4:2:0 upsampled to 4:2:2) even if the luma is isn't.
The hardware handling of packed bframes was always questionable but
it used to ok with my workaround. Today, not so much. But today we
have a bsf to unpack the bframes, so let's just use that and be
done with it.
With all the various refactorings that have happened over the years,
the current pts logic is very broken for non-trivial cases (ie: ones
where not every frame/field has a meaningful pts assocated with it).
Generally, we do not want to write AV_NOPTS_VALUE as the output
timestamp, regardless of anything else. It's better to pass zero
if there's no other information.
Additionally, interlaced content where the decoder returns each field
separately can result in the first field carrying the timestamp and
the second having AV_NOPTS_VALUE. It's clearly wrong to overwrite
the valid timestamp.
So, let's just never write AV_NOPTS_VALUE into an output frame.
Empirically, this fixed playback of interlaced mpeg2 and h.264 and
mpeg4-asp with packed b-frames in an avi container.
Although the old API is supposed to be functional, the crystalhd
decoder is currently not working for non-annex.b h.264 content.
So, let's update to the modern API and make it work again.
Signed-off-by: Philip Langdale <philipl@overt.org>
Istvan Sebok provided a sample where field pair -> two fields content
was being misdetected by the existing logic. I added an additional
test to check the input picture type as identified by our h.264
parser.
Signed-off-by: Philip Langdale <philipl@overt.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
With the flag in place, it's hard to actually use the decoder, and
I'm happy with how it works, with the exception of DivX3 where I've
never found a sample that worked that I was confident actually
matched what the hardware claimed to support.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This was a regression that came in when I switched to using the
h.264 annex b filter all the time. As the filter modifies extradata,
its use violates the statelessness assumption that exists in the
'ffmpeg' command line tool, and maybe elsewhere. It assumes that
a docoder can be reinitalised and pointed to an existing stream and
get the same results.
For now, the only way to meet this requirement is to backup the
extradata.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Now that we're converting all streams to Annex B format, we
can identify them as such to the hardware.
Signed-off-by: Philip Langdale <philipl@overt.org>
As we're now always running mp4 format streams through the annex b
filter, it makes sense to pass the filtered stream down, as
libcrystalhd would be doing the conversion internally anyway.
Signed-off-by: Philip Langdale <philipl@overt.org>
Originally, we needed to restore the original extradata after
initialising the mp4toannexb filter because mplayer would end up
taking two passes through the init sequence for the same stream
and end up miscategorising the stream. This doesn't seem to happen
anymore, making the backup/restore process unnecessary.
Signed-off-by: Philip Langdale <philipl@overt.org>
The H.264 parser that we use to detect interlacing can only handle
an Annex B stream, so we need to actually use the filter. This is
unfortunate as the crystalhd library is already doing this conversion
internally. A future change will reorganise the decode path more
completely so that we can feed the converted stream into libcrystalhd
and avoid the second conversion.
Signed-off-by: Philip Langdale <philipl@overt.org>
In preparation for using the filter on the actual bitstream, we need
to extend it's lifetime to match that of the decoder.
Signed-off-by: Philip Langdale <philipl@overt.org>
I still don't fully understand the cause but the difference between
the samples that trigger the bug and the samples that don't is
that the former uses delay frames and the later uses drop frames
as placeholders for the packed frame. So, if we see the one type
of frame, we can assume the bug will or won't be present.
Right now, I'm detecting the frame types by size, which may not be
safe in general, but given the specific codec and file type, I
expect any scenario where we encounter these frames where they
aren't being used for b-frame packing won't care one way or
another whether the work around is in effect or not.
Signed-off-by: Philip Langdale <philipl@overt.org>
The CrystalHD hardware can do scaling, which is particularly
desirable when dealing with some high resolution clips that take
so long to decode and copy out that they end up playing back
slower than realtime. By using scaling, we can make the output
frames smaller and reduce the copy out time.
This option takes the desired horizontal width in pixels, and
the hardware will do an aspect-scale. Upscaling is not supported
and the hardware will simply ignore any request to do so.
Signed-off-by: Philip Langdale <philipl@overt.org>
James Almer
Andreas Rheinhardt
wm4
Philip Langdale
Muhammad Faiz
Michael Niedermayer
Clément Bœsch
Clément Bœsch
Paul B Mahol
Clément Bœsch
Lou Logan
sebist
Philip Langdale