It's not enough to have timeout or event based VPS/SPS/PPS information
sent in RTP packets. There are some scenarios when key frames may appear
more frequently than once a second, in which case the minimum timeout
for "config-interval" of 1 second for sending VPS/SPS/PPS isn't enough.
It might also be desirable in general to make sure the VPS/SPS/PPS is
available with every keyframe (packet loss aside), so receivers can
actually pick up decoding immediately from the first keyframe if
VPS/SPS/PPS is not signaled out of band.
This commit adds the possibility to send VPS/SPS/PPS with every key frame.
This mode can be enabled by setting "config-interval" property to -1. In
this case the payloader will add VPS, SPS and PPS before every key (IDR)
frame.
https://bugzilla.gnome.org/show_bug.cgi?id=757892
The MPEG standard (ISO-13880-1) says the reserve bits need to be set
to one (2.1.64). This is causing transport streams to fail validation
on some systems.
https://bugzilla.gnome.org/show_bug.cgi?id=760127
This works usually in this place, unless the compiler optimizes things in
interesting ways in which case it causes stack corruption and crashes later.
The compiler in question here is clang with -O1, which seems to pack the stack
a bit more and causes writing to the guint as pointer to overwrite map.memory,
which then later crashes during unmapping of the memory.
While encoding the frame in ASCII mode, per component four bytes are needed
and after every 20 bytes, a \n will be added. So the calculation should be
size = size * (4 + 1 / 20). This should exclude the header being written.
Since header is also being included in the calculations, memory mishandlings
are happening.
https://bugzilla.gnome.org/show_bug.cgi?id=759520
rename gst-launch --> gst-launch-1.0
replace old elements with new elements(ffmpegcolorspace -> videoconvert, ffenc_** -> avenc_**)
fix caps in examples
https://bugzilla.gnome.org/show_bug.cgi?id=759432
The edit rate is only supposed to be the same in a source package, but there
might be multiple source packages with the same essence container. As such
just comparing the body/index SID is not sufficient.
This was completely broken before and could only work on a very constrained
set of files. After these changes it should work except for situations where
PTS != DTS, which is not handled at all in mxfdemux currently.
https://bugzilla.gnome.org/show_bug.cgi?id=759118
According to S377-1-2009c 9.2 the local tags must not be resolved from the
primer pack, which as a result means that there can't be any other tags than
statically assigned ones.
This is to support byte-stream decoder that does not remember the
PPS/SPS after a flush. This is not needed by all decoders, but is
harmless for those that do remember.
https://bugzilla.gnome.org/show_bug.cgi?id=758405
The order in which program switch must happen is:
1) drain all data on old pads (but don't push EOS)
2) add new pads (but don't push any data on them)
3) Push EOS and remove old pads
4) Start pushing data on new pads
There was one caveat in this implementation, which is that when
we activate a sparse pad (step 2) we would push a GAP event. The problem
is that, while being an event, it is actually *data*.
We therefore need to make sure pushing those GAP event is done at the step
we start pushing data.
https://bugzilla.gnome.org/show_bug.cgi?id=750402
Before we add any streams, make sure we drain all streams. This ensures
there's consistency that only "new" data will be pushed on buffers once
the new pads are added
https://bugzilla.gnome.org/show_bug.cgi?id=750402
When changing programs, the order of events needs to be the following:
* add pads from new program
* send EOS on old pads
* remove old pads
* emit 'no-more-pads'
Previously tsdemux was not doing that, and was first deactivating and
removing old pads before adding new ones.
We fix this by allowing subclasses of mpegtsbase to be able to handle
themselves the deactivation of programs. In this case tsdemux will
properly deactivate it once it has activated the new program.
https://bugzilla.gnome.org/show_bug.cgi?id=750402
The videoframe-audiolevel element acts like a synchronized audio/video "level"
element. For each video frame, it posts a level-style message containing the
RMS value of the corresponding audio frames. This element needs both video and
audio to pass through it. Furthermore, it needs a queue after its video
source.
https://bugzilla.gnome.org/show_bug.cgi?id=748259
0x04 signifies a MPEG elementary stream but according to RP2008, 0x10 should
be used for a h264 byte-stream. This also fixes compatibility of our files
with ffmpeg.
If packet->payload_unit_start_indicator is true and pointer 0, there is no
discontinuity check. Therefore there could be a previous section not complete
that need to be cleared.
https://bugzilla.gnome.org/show_bug.cgi?id=758010
The values of channel_mapping are copied by gst_codec_utils_opus_create_caps ()
but it doesn't free or take ownership of the g_new0 allocated memory. This
needs to be freed before going out of scope.
CID 1338692
buf surely isn't NULL inside the block conditional to a buffer size bigger
than (G_MAXUINT16 - 3). Plus gst_buffer_unref() checks if the buffer is
NULL and does nothing if it is.
CID 1338693
If tsdemux never receives data for a stream, the corresponding pad will never
be added and stream->active will remain FALSE. When the stream is removed, the
pad will not be unreffed and will be leaked.
https://bugzilla.gnome.org/show_bug.cgi?id=757873
The current implementation for detecting the resolution changes
on key frames is based on vp8 bitstream alignment. Avoid this
width and height parsing for vp9 bitstream, which requires proper
frame header parsing inorder to detect the resolution change (Fixme).
https://bugzilla.gnome.org/show_bug.cgi?id=757825
It is up to the element handling the seek to send flush events
downstream, otherwise we end up with a situation where upstream
would get unexpected GST_FLOW_FLUSHING
The Onvif Streaming Specification specifies that the NTP timestamps
in the Onvif extension header indicaes the absolute UTC time associated
with the access unit. But by using running time we can not achieve that,
since a frame's running time depends on the played interval, whether a
non-flushing is done, etc. Instead we have to use the stream time.
https://bugzilla.gnome.org/show_bug.cgi?id=757688
It is now possible to update the currently used ntp-offset with a
custom serialized downstream event. The element will read the ntp-offset
property when doing the state transition from READY to PAUSED and
use that offset until it receives a "GstNtpOffset" event, which also
has a "ntp-offset" attribute in that it's structure. In case the
property is not set and no event has been received, the element will
guess the npt-offset with help of the clock. If no clock can be
retrieved, the element will error out and stop the data flow.
The same event is also used for updating the D/E-bits in the RTP
extension header. The discont flag in a buffer can be set whenver a
live/network source looses a frame, but that is not the type of
discontinuity that the onvif extension header should reflect. The
header is mainly used for playback of a track concept, in which
gaps can be present, and it's those kind of gaps that should be
highlighted with the D- and E-bits.
https://bugzilla.gnome.org/show_bug.cgi?id=757688
If a buffer or a buffer list is cached, no events serialized with the
data stream should get through. The cached buffers and events should
be purged when we stop flushing.
https://bugzilla.gnome.org/show_bug.cgi?id=757688
Otherwise those symbols can conflict with external libraries when
linking everything statically for mobile targets.
Use the gst_gm_ prefix, short for gst geometric math.
https://bugzilla.gnome.org/show_bug.cgi?id=756882
Store and copy input state fields when setting the
output state of the decoder. Avoids problems like
the framerate set by an upstream element being ignored
Related to:
https://bugzilla.gnome.org/show_bug.cgi?id=756563
New subclass with a similar behaviour as the old liveadder, but
a slightly different API as the latency is in nanoseconds, not
milliseconds. Also, the new liveadder has a effective latency that
is latency + output-buffer-duration. In practice, just setting a non-zero
latency with the new audiomixer gives you the right behavior in 99% of the
cases.
Build error due to wrong argument type in debug message
aagg->priv->offset and next_offset are of type int64, but uint64
formatter is being used in logs. Changing all those to int64
https://bugzilla.gnome.org/show_bug.cgi?id=756065
When g_option_context_parse fails, context and error variables are not getting free'd
which results in memory leaks. Free'ing the same.
And replacing g_error_free with g_clear_error, which checks if the error being passed
is not NULL and sets the variable to NULL on free'ing.
https://bugzilla.gnome.org/show_bug.cgi?id=753854
The buffer timestamps in the collect function will already be
running time, don't try to convert them again to running time,
this would yield CLOCK_TIME_NONE now that the segment is shifted
to account for negative dts.
This fixes x264enc ! mpegtsmux ! hlssink, which was broken
because mpegtsmux would send a downstream key unit event with
running time NONE and then hlssink would immediately send
another one upstream and it would just be a flood of force
keyframe events in both directions after the first one. This
would then break hlssink because it uses multifilesink in
next-file=key-unit-event mode, and starting a new file after
every few kB does not work well for HLS.
The alsamidisrc element allows to get input event from ALSA MIDI
sequencer devices, and possibly convert them to sound using some
downstream element like fluiddec.
https://bugzilla.gnome.org/show_bug.cgi?id=738687
An IDX file or codec_data normally contains the original frame size of
the video. Allow upstream to provide this information by sending a
custom event, which will allow scaling the overlay correctly.
https://bugzilla.gnome.org/show_bug.cgi?id=663750
Instead of only supporting writing SPU data directly to YUV frames,
render the SPU data to an intermediate AYUV overlay buffer. The overlay
data is then attached to the video frame if downstream supports overlay
composition, otherwise the AYUV overlay is blended to the video frame.
For the PGS format, the overlay buffer size is set to the size of the
Composition Window, and its position in the overlay composition is set
to the window position. The objects to render are now cropped when the
cropping flag is set.
For the Vobsub format, the overlay buffer size is set to the size of the
Display Area.
Once rendered, the overlay composition rectangle is now moved and scaled
to fit the video output size, to avoid clipping.
https://bugzilla.gnome.org/show_bug.cgi?id=663750
We might've queued up a GAP buffer that is only partially inside the current
output buffer (i.e. we received it too late!). In that case we should only
skip the part of the GAP buffer that is inside the current output buffer, not
also the remaining part. Otherwise we forward this pad too far into the future
and break synchronization.
Derive from GstVideoSink so that preroll frames will automatically
get rendered too, unless the show-preroll-frame property is set to
FALSE. Fixes intervideosrc only picking up frames if intervideosink
is in PLAYING state.
https://bugzilla.gnome.org/show_bug.cgi?id=755049
Fix the negotiation of GstVideoOverlayComposition by checking
intersection with the peer caps, rather than just accept-caps,
which might only check the pad template.
https://bugzilla.gnome.org/show_bug.cgi?id=755113
We have to queue buffers based on their running time, not based on
the segment position.
Also return running time from GstAggregator::get_next_time() instead of
a segment position, as required by the API.
Also only update the segment position after we pushed a buffer, otherwise
we're going to push down a segment event with the next position already.
https://bugzilla.gnome.org/show_bug.cgi?id=753196
x/y/w/h are signed integers. As can be seen in GstCompositorPad.
The prototype for clamp_rectangle was wrong. This commit reverts the change
and fixes the prototype.
This reverts commit bca444ea4a.
CLAMP checks both if value is '< 0' and '> max'. Value will never be a negative
number since it is an unsigned integer. Removing that check and only checking if
it is bigger than max by using MIN().
CID 1320707
As it's recursive, gst_pad_get_allowed_caps() may also return
empty for anything incompatible downstream. EMPTY is not valid caps
value for gst_caps_fixate(). This lead to assertion and then crash.
Ideally, the negotiate function should be re-factored to have a return
value, and we could make the negotiation fails earlier.
https://bugzilla.gnome.org/show_bug.cgi?id=754122
The obscured check in compositor was using the dimensions of the pad to clamp
the h/w of the pad instead of the output resolution, and was doing an incorrect
calculation to do so. Fix that by simplifying the whole calculation by using
corner coordinates. Also add a test for this bug which fell through the cracks,
and just skip all the obscured tests if the pad's alpha is 0.0.
https://bugzilla.gnome.org/show_bug.cgi?id=754107
The tsdemux latency should always be added to the minimum
latency (which is always a valid clock time value). The
"cleanup" in commit a1f709c2 made it so that it would not
be added if upstream reported 0 as minimum latency (as
e.g. udpsrc would). This broke playback of live mpeg-ts
streaming in some cases, leading to playback stutter due
to a too-small configured latency for the pipeline.
https://bugzilla.gnome.org/show_bug.cgi?id=751508
In gst_live_adder_chain() function, calls to gst_buffer_copy_region() can lead
to assertion as 'offset + size <= bufsize' is not respected.
Indeed 'offset' and 'size' parameters are calculated through calling gst_live_adder_length_from_duration(),
and thus gst_util_uint64_scale_int_round().
Depending on the nearest integers, rounded values 'offset' and 'size' can then trigger the assertion.
This case mainly occurs when 'skip' value is > 0 in chain function process.
https://bugzilla.gnome.org/show_bug.cgi?id=753759
It is faster than doing a query that propagates downstream and
should be enough
Elements: faac, gsmenc, opusenc, sbcenc, voamrwbenc, adpcmenc, sirenenc
The payloader didn't copy anything so far, the depayloader copied every
possible meta. Let's make it consistent and just copy all metas without
tags or with only the video tag.
https://bugzilla.gnome.org/show_bug.cgi?id=751774
A race condition in the state change function may cause buffers to be
unreffed while they are still used by the streaming thread in
gst_rtp_h265_pay_send_vps_sps_pps() resulting in a crash. Chain up to the
parent class first in the state change function to make sure streaming
has stopped and only then free those buffers.
https://bugzilla.gnome.org/show_bug.cgi?id=741381
The SPS struct might be filled out by a call to
gst_h264_parser_parse_subset_sps, which fills out
dynamically allocated data and requires a call
to gst_h264_sps_clear() to free it. Also make sure
to clear out any allocated SPS data when returning
an error.
https://bugzilla.gnome.org/show_bug.cgi?id=753306
The payloader didn't copy anything so far, the depayloader copied every
possible meta. Let's make it consistent and just copy all metas without
tags or with only the video tag.
https://bugzilla.gnome.org/show_bug.cgi?id=751774
When playing mts files with embedded subtitles, the buffer is mapped,
but not unmapped at the end resulting in a memory leak.
Also unref event in handle_dvd_event as it takes ownership of the event.
https://bugzilla.gnome.org/show_bug.cgi?id=753539
When playing mts files with embedded subtitles, there are few event leaks.
Events are supposed to be transfer full. So if not forwarding the event,
they need to be freed.
https://bugzilla.gnome.org/show_bug.cgi?id=753539
h264parse and gstrtph264depay do the same, let's keep the behaviour
consistent. As we now include the codec_data inside the stream, this causes
less caps renegotiation.
https://bugzilla.gnome.org/show_bug.cgi?id=753228
rtph264depay does the same and this fixes decoding of some streams with 32
SPS (or 256 PPS). It is allowed to have SPS ID 0 to 31 (or PPS ID 0 to 255),
but the field in the codec_data for the number of SPS or PPS is only 5
(or 8) bit. As such, 32 SPS (or 256 PPS) are interpreted as 0 everywhere.
This looks like a mistake in the part of the spect about the codec_data.
This causes an assertion and would lead to getting a NULL instead
of a buffer. Without proper checking this would easily lead to a
segfault.
Related to rpth264depay: https://bugzilla.gnome.org/show_bug.cgi?id=737199
String parameters are expected to be passed as (f0r_param_string *),
which actually map to char**. In the filters this is evaluated as
(*(char**)param) which currently lead to crash when passing char*.
Remove the special case for string, all types, including char* as
passed as a reference.
https://phabricator.freedesktop.org/T83
Check if downstream is seekable via a SEEKING query and output a
BYTE segment if we want to seek back to fix up the headers later,
but if we're streaming send a TIME segment instead (which goes
down better with e.g. asfmux ! rtpasfpay).
https://bugzilla.gnome.org/show_bug.cgi?id=719553
Some video bitstreams report a too restrictive set of profiles. If a video
decoder was to strictly follow the indicated profile, it wouldn't support that
stream, whereas it could in theory and in practice. So we should relax the
profile restriction for allowing the decoder to get connected with parser.
https://bugzilla.gnome.org/show_bug.cgi?id=747613
In the case where you have a source giving the GstAggregator smaller
buffers than it uses, when it reaches a timeout, it will consume the
first buffer, then try to read another buffer for the pad. If the
previous element is not fast enough, it may get the next buffer even
though it may be queued just before. To prevent that race, the easiest
solution is to move the queue inside the GstAggregatorPad itself. It
also means that there is no need for strange code cause by increasing
the min latency without increasing the max latency proportionally.
This also means queuing the synchronized events and possibly acting
on them on the src task.
https://bugzilla.gnome.org/show_bug.cgi?id=745768
VPS is not mandatory, and need not check for its presence before setting
the caps. Because of the check, in streams which don't have VPS,
sticky event mishandling happens.
https://bugzilla.gnome.org/show_bug.cgi?id=752807
In media to caps function, reserved_keys array is being used for variable i,
leading to GLib-CRITICAL **: g_ascii_strcasecmp: assertion 's1 != NULL' failed
changed it to variable j
https://bugzilla.gnome.org/show_bug.cgi?id=753009
Skip keys from the fmtp, which we already use ourselves for the
caps. Some software is adding random things like clock-rate into
the fmtp, and we would otherwise here set a string-typed clock-rate
in the caps... and thus fail to create valid RTP caps
https://bugzilla.gnome.org/show_bug.cgi?id=753009
The PID on a pad shouldn't change on a state change, only
if the pad is freed and a new one created. Clearing the PID
prevented mpegtsmux from being reused, because all packets
would end up muxed in PID 0
https://bugzilla.gnome.org/show_bug.cgi?id=752999
Accumulate streamheader packets in reverse into the
GList for efficiency, and reverse the list once when
processing.
Improves muxing speed when there are a lot of
streamheaders.
Don't throw away AU delimiter(s) that precede the SPS/PPS. Should
fix MPEG-TS playback on iOS/Quicktime when muxing streams that
already have AU delimiters.
See https://bugzilla.gnome.org/show_bug.cgi?id=736213 for getting
h264parse to insert AU delimiters when they don't already
exist.
We need to sync the pad values before taking the aggregator and pad locks
otherwise the element will just deadlock if there's any property changes
scheduled using GstController since that involves taking the aggregator and pad
locks.
Also add a test for this.
https://bugzilla.gnome.org/show_bug.cgi?id=749574
ret is declared just to initialize to TRUE and overwrite with the value of
vret. We can return the value of vret directly. vret is TRUE unless the
forward_event_func sets it to FALSE.
We only want to do a hard reset of the observations if we're working
with TIME segments in push mode. For BYTE segment we want to keep
the observations (in order to do seeks in push-mode).
When in push mode, we want to discard all previous observations from the
mpegtspacketizer when we get a DISCONT buffer.
This avoids trying to calculate bogus timestamps (estimating them using old
PCR observations).
We only do a hard reset in push-mode. In pull-mode we still need the observations
(in order to seek properly)
This is not public API, use g_assert() instead of
g_return_if_fail(), so that it's compiled out in
releases. It's only called from our code, with &foo.
Introduced by c4c9fe60b pcapparse: Take buffer directly from the adapter
Using gst_adapter_take_buffer_fast() can lead to buffers that are
made up of multiple memories with the first memory smaller than the
RTP header size, which violates assumptions GstRtpBaseDepayloader
makes, namely that the complete RTP header will be in the first
memory. This leads to such packets being dropped when feeding
them from pcapparse to RTP depayloaders. Use take_buffer() so
we get buffers with a single memory.
Introduced by c4c9fe60b pcapparse: Take buffer directly from the adapter
Flush any trailing bytes after the payload from the adapter as well,
otherwise we'll read a bogus packet size from the adapter next and
then everything goes downhill from there.
https://bugzilla.gnome.org/show_bug.cgi?id=751879
Can be used to fix misbehaving sinks. It will pass through all buffers
until it encounters GST_FLOW_ERROR or GST_FLOW_NOT_NEGOTIATED (configurable).
At that point it will unref the buffers and return GST_FLOW_NOT_LINKED
(configurable) - until the next READY_TO_PAUSED or FLUSH_STOP.
https://bugzilla.gnome.org/show_bug.cgi?id=750098
Move the pixel-aspect-ratio calculations higher up in caps
determination, so the results are available for a call to
gst_video_multiview_guess_half_aspect() when stereoscopic video
is detected.
Use QOS messages to update rendered and dropped frame stats. This is
the only accurate method. The old method didn't take max-lateness and
latency into account.
After few iteration, this variable became the same as dts. It's not
the min as the name says, but the dts of the current buffer. Simply
remove and place with dts. Also move the debug trace to actually
print the signed version of the running-time dts.
after e000a6f0a4, there is build error in bad plugins
this happens because, GST_CLOCK_STIME_IS_VALID () is being checked for pad_data
but it expects a GstClockTime parameter. Changing the check to 'dts'
https://bugzilla.gnome.org/show_bug.cgi?id=750961
The segment should start at first PTS, and the vairable name lower_pts
state so correctly. Though we where using the first DTS instead. This
could lead to small desynchronization of video stream.
https://bugzilla.gnome.org/show_bug.cgi?id=740575
Use the saved DTS, make it signed and pass that to the stream muxer. This
preserves the running time sign. All usage of -1 as invalid TS are now
replaced with G_MININT64. Negative values will be seen as wrap-around
point, but the delta between PTS and DTS will remain correct. Demuxers
don't care about absolute values, they only cares about deltas.
https://bugzilla.gnome.org/show_bug.cgi?id=740575
There was code to detect backward dts, but the marker min_dts
was never set. Setting it enable this feature that prevents
potential integer overflow when generating TS.
https://bugzilla.gnome.org/show_bug.cgi?id=740575
Wait until at least one keyframe has been parsed before
deciding to switch to passthrough mode, in case the
stream contains SEI messages that supplement the output
caps - for example by providing stereoscopic information
We were off by one byte in the matching
It should be (using 24 bit matching):
* startcode : 0000 0000 0000 0000 1000 00xx
* mask (bin) : 1111 1111 1111 1111 1111 1100
* mask (hex) : f f f f f c
* match : 0 0 0 0 8 0
https://bugzilla.gnome.org/show_bug.cgi?id=750685
In case of the videomark being partially or fully outside,
an error was bein thrown saying, mark width is more than video width.
And when the width, offset properties are set to maximum it resulted in crash.
Instead of throwing error, added logic to detect the mark
in case of partial visibility or dont show the mark when it is outside.
https://bugzilla.gnome.org/show_bug.cgi?id=743908
