34af8ed66a changed the code to use the
packetizer's packets instead of the incoming buffers, but mpegtsbase
didn't actually push all packets to the subclass. As a result, padding
(PID 0x1FFF) packets got lost.
Add a new boolean to toggle pushing unknown packets to mpegtsbase and
have mpegtsparse make use of it.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1300>
tsparse leaked input buffers quite badly:
GST_TRACERS=leaks GST_DEBUG=GST_TRACER:9 gst-launch-1.0 audiotestsrc num-buffers=3 ! avenc_aac ! mpegtsmux ! tsparse ! fakesink
The input_done vfunc was passed the input buffer, which it had to
consume. For this reason, the base class takes a reference on the buffer
if and only if input_done is not NULL.
Before 34af8ed66a, input_done was used in
tsparse to pass on the input buffer on the "src" pad. That commit
changed the code to packetize for that pad as well and removed the use
of input_done.
Afterwards, 0d2e908523 set input_done
again in order to handle automatic alignment of the output buffers to
the input buffers. However, it ignored the provided buffer and did not
even unref it, causing a leak.
Since no code makes use of the buffer provided with input_done, just
remove the argument in order to simplify things a bit.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1274>
Per specification in 2.14.2 "For PES packetization, no specific data
alignment constraints apply". So we should not advertise NAL
alignment.
This bug was introduced at the same moment the alignment field was introduced
10 years ago. The plan was that alignment=none (or no alignment field) was to
be used for mpegtsdemux, but no one noticed the error. The reason is that at
the same moment, everything dealing with H264 started defaulting to AU
alignment.
https://bugzilla.gnome.org/show_bug.cgi?id=606662#c22
This patch will have a side effect that a parser is now needed after the
tsdemux element. The following pipeline will not negotiate anymore as the
mpegtsmux element requires alignment={nal,au}.
... ! tsdemux ! mpegtsmux ! ...
As a side effect, anyone that forked from tsdemux should updated their code to
fix this bug.
According to the specification, the adaptation field length must be 183 if
there is no payload data and < 183 if the packet contains an adaptation
field and payload data.
Unfortunately some payloaders always set the flag for payload data, even if
the adaptation field length is 183.
Don't return with an error in this case. Clear the payload data flag
instead and parse the adaptation field as usual. This avoids visual
artefacts for such streams.
Both 2 and 4 are supported version of AAC ADTS format stream.
So we need to set correct version to help negotiation
especially for non-autopluggable pipeline.
Some mpeg-ts (HLS, DVB, ...) streams out there have completely broken
PCR streams on which we can't reliably recover correct timestamps.
For those, provide a property that will ignore the program PCR stream
(by faking that it's not present (0x1fff)).
According to following two specs, add support for AC4 in tsdemux.
1. ETSI TS 103 190-2 V1.2.1 (2018-02) : Annex D (normative): AC-4 in MPEG-2 transport streams
2. ETSI EN 300 468 V1.16.1 (2019-08) : Annex D (normative):Service information implementation of AC-3, EnhancedAC-3, and AC-4 audio in DVB systems
alignment works like in mpegtsmux, joining several MpegTS packets into
one buffer. Default value of 0 joins as many as possible for each
incoming buffer, to optimise CPU usage.
... by seeking to target offset determined by new seek segment,
rather than that of the previous segment. The latter would typically
seek back to start for a non-accurate seek, and lead to a lot
of skipping in case of an accurate seek.
Currently tsdemux timestamps only the PTS, and only issues the DTS if
it's different. In that case, parsers tend to estimate the next DTS
based on the previous DTS and the duration, which can accumulate
rounding errors.
Packets of a given PID are meant to have sequential continuity counters
(modulo 16). If there are not sequential, this is the sign of a broken
stream, which we then consider as a discontinuity.
But if that new packet is a frame start (PUSI is true), then we can resume
from that packet without any damage.
This went un-noticed for 6 years :( The issue is that for short
sections (without subtables and CRC), we would always fail when
checking whether we had enough data or not and then default to the
long section checking.
Use the long section checking would then cause interesting side-effects
for short sections (such as believing they were already seen and therefore
would be dropped/ignored).
Allows for "low latency" mpeg-ts mode which is not standard, but somewhat common.
For this to work the sender has to put timestamps at a higher frequency than the spec requires.
PES packets with size 0 are unbounded, and
could therefore overflow the 32-bit size
accumulator.
Add a 32MB limit, which is larger than
any PES packet should ever get. If one does,
then output a 32MB chunk and continue.
Don't signal a pipeline error when processing incomplete
j2pk PES packets that are too small. That can happen normally
during a DISCONT and shouldn't shut down the whole pipeline
Remove some custom and incomplete seek calculation
logic in favour of gst_segment_do_seek(), and
short-circuit any actual seeking or recalculation
if the position didn't change and just send an updated
segment directly.
This removes the custom seeking logic in favour of
using standard core seek handling.
The MPEG-TS packetiser should use the upstream DTS for
skew correction when running in that mode, as the DTS
carries the upstream arrival time. The PTS (if it's
set at all) is less useful, and can be invalid.
Unless we only have sparse streams. In this case we will consider them.
It fixes a bug happening when first observed timestamp comes from a
sparse stream and other streams don't have a valid timestamp, yet. Thus
leading the timestamp from sparse stream to be the start of the
following segment. In this case, if the timestamp is really bigger than
non-sparse stream (audio/video), it will lead the pipeline to clip
samples from the non-parse stream.
https://bugzilla.gnome.org/show_bug.cgi?id=744469
* Avoid copying the pending data and instead create a buffer directly from
that data with the appropriate offset.
* Locate the jp2k magic to determine the exact location of the (first) frame
data instead of assuming that the header is of an expected size
https://bugzilla.gnome.org/show_bug.cgi?id=786111
The jp2k specification (ITU-T T.800) specifies that the 'brat' box
has two fields and the second one (AUF2) can be set to 0 for progressive
streams.
The problem is that the mpeg-ts specification (ITU-T H.222.0 06/2012)
says that the AUF2 field is only present if the stream is interlaced
In order to cope with both situation, accept those next 32bit if the
stream is marked as progressive and those bits contain 0
https://bugzilla.gnome.org/show_bug.cgi?id=786111
Doing lazy conversion of PCR values doesn't work right
when a PCR discont is encountered. Instead, convert PCR
values to the continuous timestamp domain as soon as we
encounter them and store that instead.
