This way, the caller doesn't need to coordinate setting the option
after the moov atom has been written. The downside is that it is
no longer possible to use the option for checking whether the moov
atom already has been written, but a caller is able to keep track
of that by other means anyway.
Signed-off-by: Martin Storsjö <martin@martin.st>
The previous use of the mov->fragments field, for determining whether
written packets were part of the first fragment or not, didn't
work as intended when using the empty_moov flag.
Signed-off-by: Martin Storsjö <martin@martin.st>
By making sure we at each time only have one pointer set, either a
local variable or one in the context, we avoid potential double frees
in the cleanup routines. If chain->rtp_ctx is set, it is closed by
calling avformat_write_trailer, but that shouldn't be called unless
avformat_write_header succeeded.
This issue was pointed out by Andreas Cadhalpun.
Signed-off-by: Martin Storsjö <martin@martin.st>
In this case len is always at least 3, since it is checked against
RTP_HEVC_PAYLOAD_HEADER_SIZE + 1 before entering the switch block.
Bug-Id: CID 1238784
This avoids assuming that e.g. audio samples are marked as
sync samples.
This allows omitting the sample flags from trun, if the default
flags happen to be right for all the samples.
Signed-off-by: Martin Storsjö <martin@martin.st>
Use correct context, reduce log level, don't assume it is a video stream,
and print the tag of the unknown stream.
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
a876585215 had the unintended side effect of returning AVERROR(ENOMEM)
when track->entry is zero, while the code intentionally wants to
continue in that case.
Signed-off-by: Martin Storsjö <martin@martin.st>
The mov muxer already supports picking up extradata that wasn't
present during the avformat_write_header call - we just need to
propagate it. Since the dash muxer uses delay_moov, we have time
up until the first segment is written to get extradata filled in.
Also update the codec description string when the extradata becomes
available.
Signed-off-by: Martin Storsjö <martin@martin.st>
It is used in adx_read_packet, which currently depends on the
decoder/parser setting this value between reading the file header and
demuxing the first packet.
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
The chunk size is limited to UINT16_MAX (written by avio_wb16), so make
sure that the packet size is not too large.
Such large frames need to be split into slices smaller than 64 kB, but
that is currently supported neither by the rv10/rv20 encoders nor the rm
muxer.
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
The original flags variable contains rtpdec flags, while the
rmflags variable contains RM flag bits which have a completely
different definition.
Signed-off-by: Martin Storsjö <martin@martin.st>
The only case where RTP_FLAG_KEY actually is needed is
in RDT, where such a flag needs to be passed via the
rtpdec parse function's flags parameter.
Signed-off-by: Martin Storsjö <martin@martin.st>
Nothing in the framework nor in the rest of the depacketizer actually
uses this flag - the chained demuxer sets the keyframe flag properly on
demuxed packets already.
Signed-off-by: Martin Storsjö <martin@martin.st>
This fixes an oversight in 96084251, in a refactoring done on top
of Gilles' original patch.
Pointed out by Gilles Chanteperdrix.
Signed-off-by: Martin Storsjö <martin@martin.st>
Similarly to what has been done for MOV, display XMP metadata only when
users explicitly require it.
The Extensible Metadata Platform tag can contain various kind of data
which are not strictly related to the video file, such as history of
edits and saves from the project file.
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
This reverts commit 4abfa387b8.
This commit broke playback of fragmented mp4 files with b-frames.
While investigating this, it turned out that the general framework
isn't ready for a PTS-based index yet. Revert this change until
a better thought out solution is in place.
Signed-off-by: Martin Storsjö <martin@martin.st>
Instead check the timestamps while muxing, to avoid buffering a
too long timestamp range into one single packet.
This makes the AMR and AAC packetization slightly less efficient,
since we set a possibly unnecessarily high max_frames_per_packet.
(These packetizers end up doing a memmove of the TOC bytes if
sending a packet before max_frames_per_packet is achieved, and
we end up setting max_frames_per_packet to a value that should
be high enough for most uses.)
All packetizers that use max_frames_per_packet now set it either
to a default value, or to a value calculated based on other
parameters, so none of them rely on the previous default setting.
For iLBC, copy one frame at a time, to allow checking the timestamp
range for each of them - basically doing potentially multiple
loops to simplify the code instead of trying to calculate the
number of frames to buffer while honoring s1->max_delay.
This is in preparation for reducing the coupling between libavformat
and libavcodec, by not having the muxers use the encoder field
frame_size (which may not be available during e.g. stream copy).
Signed-off-by: Martin Storsjö <martin@martin.st>
Factorize out the s->num_frames check at the start of the if statements,
simplifying adding more alternative causes for sending the buffered
frames.
Signed-off-by: Martin Storsjö <martin@martin.st>
This avoids having to jump to the defaultcase in the switch. Manually
override the stream time base back to 90 kHz for the few audio codecs
that don't use the sample rate as time base (mp2, mp3).
Signed-off-by: Martin Storsjö <martin@martin.st>
This doesn't fix any bug, but makes the code simpler for later
patches, and more straightforward to read as is.
Signed-off-by: Martin Storsjö <martin@martin.st>
After sending a fragmented frame, len (s->buf_ptr - s->buf) isn't
zero, while s->num_frames is zero as intended. Using s->num_frames
makes it work as intended, and is less convoluted than keeping track
of (resetting) s->buf_ptr.
This avoids sending stray data after sending a fragmented aac packet.
CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
Don't prefix them ffio_url, which is misleading, sounding too
much like the urlprotocol layer (like ffurl_*).
Signed-off-by: Martin Storsjö <martin@martin.st>
Many of these functions were named foo_free_context, and since
the functions no longer should free the context itself, only
allocated elements within it, the previous naming was slightly
misleading.
Signed-off-by: Martin Storsjö <martin@martin.st>
This is different from how it is handled in codecs/demuxers/muxers
though (where the close function isn't called if the open function
failed), but since the number of depacketizers that have an .init
function is quite limited, this is easy to change.
The main point is that if the init function failed, we shouldn't
try to use that depacketizer at all - this makes sure that the
parse function doesn't need to check for the things that were
initialized in the init function.
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes it more consistent with depacketizers that don't have any
.free function at all, where the payload context is freed by the
surrounding framework. Always free the context in the surrounding
framework, having the individual depacketizers only free any data
they've specifically allocated themselves.
This is similar to how this works for demuxer/muxers/codecs - a
component shouldn't free the priv_data that the framework has
allocated for it.
Signed-off-by: Martin Storsjö <martin@martin.st>
They share a great deal of common structure; only a few minor
bits in the headers differ.
This also fixes an off-by-one in sending of the last fragment
of large HEVC nals (where it previously sent len+2 bytes, even
if it should have been len+RTP_HEVC_HEADERS_SIZE aka len+3).
Signed-off-by: Martin Storsjö <martin@martin.st>
This allows getting rid of quite a bit of boilerplate in depacketizers.
The default value (initializing need_parsing to 0, aka
AVSTREAM_PARSE_NONE) is the same as it is initialized to by default
in AVStream.
Signed-off-by: Martin Storsjö <martin@martin.st>
When a client behind a NAT issues a pause command, and stay paused for a
long time, the router may stop the RTP/RTCP port redirection. Resend the
hole punching packets before each PLAY command to cause the router to
restart the port redirection in that case.
Move the existing code for sending the packets from the SETUP phase
to the PLAY phase.
Signed-off-by: Martin Storsjö <martin@martin.st>
When receiving an RTCP packet, the difference between the last RTCP
timestamp and the base timestamp may be negative. As these timestamps
are of the uint32_t type, the result becomes a large integer. Cast
the difference to int32_t to avoid this issue.
The result of this issue is very large start times for RTSP
streams, and difficulty to restart correctly after a pause.
Signed-off-by: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
Signed-off-by: Martin Storsjö <martin@martin.st>
If src_len is too small for nal_size, we already print a warning
above, and the next step is to check the while loop condition
anyway, so this one serves no purpose.
Signed-off-by: Martin Storsjö <martin@martin.st>
Previously, errors were only logged but the code kept on trying,
and never actually returning the error as a return value.
Signed-off-by: Martin Storsjö <martin@martin.st>
Including libavcodec/get_bits.h is superfluous for AV_RB16 - nothing
in this file uses any actual bitstream reader.
Signed-off-by: Martin Storsjö <martin@martin.st>
This allows the output to be used with stream copy, which discards
packet from the start until the first keyframe.
Signed-off-by: Martin Storsjö <martin@martin.st>
During remuxing avcodec_copy_context() is discouraged as certain fields
(such as codec_tag) could reflect invalid values between input and
output contextes.
CTS-based seek is reasonable since player requests frames in output order
not coded order.
This change fixes seek to a keyframe within consecutive keyframes.
Let's say P[0|-1] and P[1|0], here x and y inside [x|y] are PTS and DTS
respectively, and both two frames are a keyframe. If you try to seek on
PTS=0, i.e. P[0|-1], you'll get P[1|0] if the demuxer is DTS based. This
is obviously undesirable.
Signed-off-by: Martin Storsjö <martin@martin.st>
analyze() is currently called both when probing and from read_header().
It determines the packet start by looking for the sync byte, followed by
unset Transport Error Indicator and valid adaptation_field_control.
This makes sense to do when probing, but once we already know the format
is MPEG-TS, it is counterproductive to be so strict -- e.g. in some
files the TEI might be set and analyze() might get called with a smaller
buffer than the one used for probing, resulting in a failure.
Nothing uses it, and it provides no public API.
Archeological finds:
Commit 101036adb9 added the API.
Commit a8dd8dc6e9 made mpegts.c use it.
Commit af8aae3fa3 disabled it by default in mpegts.c.
Commit ae2bb52cd2 removed all uses of this from mpegts.c.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
This atom typically is used for a track title. The handler name is stored
as a Pascal string in the QT specs (first byte is the length of the string),
so do not export it.
A second length check based on the first character is added to avoid
overwriting an already specified handler_name (it happens with YouTube
videos for instance, the handler_name get masked), or specifying an
empty string metadata.
The Pascal string fix and the second length check are written
by Clément Bœsch <clement.boesch@smartjog.com>.
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Trigger a refill if the seek action moves the pointer
at the end of the buffer.
Before this patch the read action following the seek would trigger
the refill, while write action would write outside the buffer.
In the Libav codebase few muxers seek forward outside of what
already has been written so it is quite unlikely to experience
the problem with the default buffer size.
CC: libav-stable@libav.org
This partially reverts cf70ba37ba, since
it didn't take into account when rotation is 0, but there is another
valid operation (eg. translation) in the matrix.
Found-by: Michael Niedermayer <michaelni@gmx.at>
This goto wasn't necessary originally, but it should have been
added when the write_manifest call was added in 8e276378.
CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
This fixes sending chunked packets (packets larger than the output
chunk size, which often can be e.g. 4096 bytes) with a timestamp delta
(or absolute timstamp, if it's a timestamp step backwards, or the
first packet of the stream) larger than 0xffffffff.
The RTMP spec explicitly says (in section 5.3.1.3.) that packets of
type 3 (continuation packets) should include this field, if the
previous non-continuation packet had it included.
The receiving code handles these packets correctly.
Pointed out by Cheolho Park.
CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
When the display matrix is not the identity one, but the rotation angle
is zero, there is no need to update the sample aspect ratio.
Otherwise, it is possible to obtain negative values which interferes
with transcoding in later stages. This kind of behaviour is reproducible
on mov files with "major_brand: MSNV".
CC: libav-stable@libav.org
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
A failure in segment_end() or segment_start() would lead to freeing
a dangling pointer and in general further calls to seg_write_packet()
or to seg_write_trailer() would have the same faulty behaviour.
CC: libav-stable@libav.org
Reported-By: luodalongde@gmail.com
This comment can be traced back to the initial commit from 2001,
and it seemed to be misleading/incorect already back then. (It
was used for normal, non-raw file formats already then.)
Signed-off-by: Martin Storsjö <martin@martin.st>
This should be more correct. This also should give more sensible
switching between video streams with different amount of b-frame
delay.
The current dash.js release (1.2.0) fails to start playback of
such files from the start (if the start pts is > 0), but this has
been fixed in the current git version of dash.js.
Also enable the use of edit lists, so that streams in many cases
start at pts=0.
Signed-off-by: Martin Storsjö <martin@martin.st>
Use the more generic approach with the delay_moov flag, instead of
having a update mechanism specific to this one single atom.
Signed-off-by: Martin Storsjö <martin@martin.st>
This delays writing the moov until the first fragment is written,
or can be flushed by the caller explicitly when wanted. If the first
sample in all streams is available at this point, we can write
a proper editlist at this point, allowing streams to start at
something else than dts=0. For AC3 and DNXHD, a packet is
needed in order to write the moov header properly.
This isn't added to the normal behaviour for empty_moov, since
the behaviour that ftyp+moov is written during avformat_write_header
would be changed. Callers that split the output stream into header+segments
(either by flushing manually, with the custom_frag flag set, or by
just differentiating between data written during avformat_write_header
and the rest) will need to be adjusted to take this option into use.
For handling streams that start at something else than dts=0, an
alternative would be to use different kinds of heuristics for
guessing the start dts (using AVCodecContext delay or has_b_frames
together with the frame rate), but this is not reliable and doesn't
necessarily work well with stream copy, and wouldn't work for getting
the right initialization data for AC3 or DNXHD either.
Signed-off-by: Martin Storsjö <martin@martin.st>
If fragments == 0 it means we haven't written any moov atom yet.
If the empty_moov flag is set, we already have written an empty moov
atom at startup. Thus, the check for empty_moov is redundant.
This is in preparation for allowing writing the moov atom later,
even when using the empty moov flag.
Signed-off-by: Martin Storsjö <martin@martin.st>
When writing an explicit time, reset the cur_time variable to this
value as well. This avoids writing excessive time attributes for each
segment in the timeline, as long as the segments are continuous.
Signed-off-by: Martin Storsjö <martin@martin.st>
The MoveFileExA is available in the headers regardless which API
subset is targeted, but it is missing in the Windows Phone link
libraries. When targeting Windows Store apps, the function is
available both in the headers and in the link libraries, and thus
there is no indication for the build system that this function
should be avoided - such an indication is only given by the
Windows App Certification Kit, which forbids using the MoveFileExA
function.
Therefore check the WINAPI_FAMILY defines instead, to figure out
which API subset is targeted.
Signed-off-by: Martin Storsjö <martin@martin.st>
Since the mpegts muxer now can handle being called with a NULL
AVIOContext, we don't need to try to allocate one before calling
write_trailer.
Signed-off-by: Martin Storsjö <martin@martin.st>
If opening and closing dynamic buffers as AVIOContext, we may
not have any AVIOContext available when wanting to close and
deallocate the muxer. Allow calling write_trailer despite this.
Signed-off-by: Martin Storsjö <martin@martin.st>
The pts and the corresponding duration is written in sidx
atoms, thus make sure these match up correctly.
Signed-off-by: Martin Storsjö <martin@martin.st>
Since this structurally is quite different from normal RTP
(multiple streams are muxed into one single mpegts stream,
which is packetized into one single RTP session), it is kept
as a separate muxer.
Since this structurally also behaves differently than normal
RTP, all of the other muxers that do chained RTP muxing
(rtsp, sap, mp4) would need to be updated similarly to handle
this - in particular, creating one single rtp_mpegts muxer
for the whole presentation instead of one rtp muxer per stream.
Signed-off-by: Martin Storsjö <martin@martin.st>
The packetizer only supports splitting at GOB headers - if
such aren't available frequently enough, it splits at any
random byte offset (not at a macroblock boundary either, which
would be allowed by the spec) and sends a payload header pretend
that it starts with a GOB header.
As long as a receiver doesn't try to handle such cases cleverly
but just drops broken frames, this shouldn't matter too much
in practice.
Signed-off-by: Martin Storsjö <martin@martin.st>
Instead explicitly jump to the default case in the cases where
it is wanted, and avoid fallthrough between different codecs,
which could easily introduce bugs if people editing the code
aren't careful.
Signed-off-by: Martin Storsjö <martin@martin.st>
If we throw away the buffered incomplete frame, make sure to also
throw away the buffered bits of an incomplete byte at the same
time.
CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
In particular, when packetizing mpegts into rtp, the input packet
timestamp may come from more than one stream, which could cause
multiple packets be written with the same timestamp.
Signed-off-by: Martin Storsjö <martin@martin.st>
This is the same adjustment that the mp4 muxer does to the start
timestamp of fragments, since the timestamp of a sample in an mp4
file is implicit from the sum of earlier sample durations.
This avoids gaps in the timeline (which can stop dash.js from
playing it back), and makes sure the timestamp on the segmenter
level matches what the mp4 muxer actually writes into the segments.
This is only an issue if the AVPacket duration of the last
packet of a segment doesn't point to the actual start timestamp
of the next packet (the first in the next segment).
Signed-off-by: Martin Storsjö <martin@martin.st>
Write a new start time if the duration of the previous segment
didn't match the start of the next one. Check that segments
actually are continuous before writing a repeat count.
This makes sure timestamps deduced from the timeline actually
match the real start timestamp as written in filenames (if
using a template containing $Time$).
Signed-off-by: Martin Storsjö <martin@martin.st>
Since 3cec81f4d4, a zero-length metadata value would try to
allocate 2*0 bytes, where av_malloc() returns NULL.
Always add one to the allocated length, to allow space for
a null terminator in the zero-length case.
Incidentally, this fixes fate-alac on RVCT 4.0, where a compiler
bug seems to mess up the mov muxer to the point that it writes
the wrong sort of metadata. Previously this bug was undetected,
but since 3cec81f4d4 such mov files started returning
AVERROR(ENOMEM) in the mov demuxer.
Signed-off-by: Martin Storsjö <martin@martin.st>
In matroska_read_seek(), |tracks| is assigned at the begining of the
function. However, functions like matroska_parse_cues() could reallocate
the tracks and invalidate |tracks|.
This assigns |tracks| only before using it, so that it will not get
invalidated elsewhere.
Bug-Id: chromium/427266
For the last_duration field, it's mostly theoretical, but the
total_duration field more probably may need to actually be 64 bit.
Bug-Id: CID 1254944
Signed-off-by: Martin Storsjö <martin@martin.st>
As the manifest/segments are flushed to disk, log to stderr the
progress, when in verbose logging mode
Signed-off-by: Martin Storsjö <martin@martin.st>
Only the upper 2 bits of the first byte are known to be
a fixed value.
The lower bits in the first byte of a RTP packet could be set
if the input is from another RTP packetizers than libavformat's,
but for RTCP packets, they would also be set when sending RTCP RR
packets, triggering false warnings about incorrect input format
to the protocol.
Signed-off-by: Martin Storsjö <martin@martin.st>
The RTP muxer enables the actual codepaths within sdp.c,
which depend on hevc.o since e5cfc8fd.
This fixes builds with --disable-everything --enable-muxer=rtp.
Signed-off-by: Martin Storsjö <martin@martin.st>
