Commit graph

8 commits

Author SHA1 Message Date
Ederson de Souza
7edaeb3fae avtpcvfpay: Warn about timestamp issues on non-flushing seek
Seek events will cause new segments to be sent to avtpcvfpay, and for
flushing seeks, a pipeline running time reset. This running time
reset, which effectively changes pipeline base time, will cause
avtpcvfpay element to generate incorrect DTS for the initial set of
buffers sent after FLUSH_STOP.

This happens due the fact that base time change happens only when the
sink gets the first buffer after the FLUSH_STOP - so avtpcvfpay used
the wrong base time to do its calculations.

However, if the pipeline is paused before the seek, sink will update
base time when pipeline state goes to PLAYING again, before avtpcvfpay
gets the first buffers after the flush. Then avtpcvfpay element will be
able to normally calculate DTS for the outgoing packets.

This patch simply adds a warning message in case a flushing seek is
performed on a playing pipeline.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1004>
2020-05-02 17:42:15 +00:00
Ederson de Souza
12838af353 avtpcvfpay: Ensure NAL fragments are transmitted following stream specs
TSN streams are expected to send packets to the network in a well
defined "pace", which is arbitrarily defined for each stream. This pace
is defined by the "measurement interval" property of a stream.

When the AVTP CVF payloader element - avtpcvfpay - fragments a video
frame that is too big to be sent to the network, it currently defines
that all fragments should be transmitted at the same time (via DTS
property of GstBuffers generated, as sink will use those to time the
transmission of the AVTPDU). This doesn't comply with stream definition,
which also has a limit on how many packets can be sent on a given
measurement interval.

This patch solves that by spreading in time the DTS of the GstBuffers
containing the AVTPDUs. Two new properties, "measurement-interval" and
"max-interval-frames", added to avptcvfpay element so that it knows
stream measurement interval and how many AVTPDUs it can send on any of
them. More details on the method used to proper spread DTS/PTS according
to measurement interval can be found in a code commentary inside this patch.

Tests also added for the new property and behaviour.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1004>
2020-05-02 17:42:15 +00:00
Ederson de Souza
f1976e0de5 avtp: Plug several leaks
After finally running tests with valgrind enabled, some leaks were found
- both on code and on tests themselves. This patch plugs them all!
2020-02-07 21:53:57 +00:00
Ederson de Souza
c45c235b2a avtpcvfpay: Do not hide or modify gst_pad_push errors
Current code would change any non-ok return from gst_pad_push to
GST_FLOW_ERROR, thus hiding meaningful returns such as GST_FLOW_EOS.

Tests also added.
2019-11-19 13:35:00 +00:00
Ederson de Souza
416afd5fa9 avtp: CVF - fix error message
The error is about *not* being able to map a buffer.
2019-07-30 11:34:31 -07:00
Ederson de Souza
f9a16731d1 avtp: CVF - Do not infinite loop trying to fragment zero sized NAL unit
Zero sized NAL-units should not happen, but if they do, do not infinite
loop. Added also a unit test for this case.
2019-07-30 11:34:31 -07:00
Ederson de Souza
b056297eea avtp: Add fragmentation feature to CVF payloader
Based on `mtu` property, the CVF payloader is now capable of properly
fragmenting H.264 NAL units that are bigger than MTU in several AVTP
packets.

AVTP spec defines two methods for fragmenting H.264 packets, but this
patch only generates non-interleaved FU-A fragments.

Usually, only the last NAL unit from a group of NAL units in a single
buffer will be big enough to be fragmented. Nevertheless, only the last
AVTP packet sent for a group of NAL units will have the M bit set (this
means that the AVTP packet for the last fragment will only have the M
bit set if there's no more NAL units in the group).
2019-07-03 09:59:35 -07:00
Ederson de Souza
3b4f3a0b3f avtp: Introduce AVTP CVF payloader element
This patch introduces the AVTP Compressed Video Format (CVF) payloader
specified in IEEE 1722-2016 section 8. Currently, this payload only
supports H.264 encapsulation described in section 8.5.

Is also worth noting that only single NAL units are encapsulated: no
aggregation or fragmentation is performed by the payloader.

An interesting characteristic of CVF H.264 spec is that it defines an
H264_TIMESTAMP, in addition to the AVTP timestamp. The later is
translated to the GST_BUFFER_DTS while the former is translated to the
GST_BUFFER_PTS. From AVTP CVF H.264 spec, it is clear that the AVTP
timestamp is related to the decoding order, while the H264_TIMESTAMP is
an ancillary information to the H.264 decoder.

Upon receiving a buffer containing a group of NAL units, the avtpcvfpay
element will extract each NAL unit and payload them into individual AVTP
packets. The last AVTP packet generated for a group of NAL units will
have the M bit set, so the depayloader is able to properly regroup them.

The exact format of the buffer of NAL units is described on the
'codec_data' capability, which is parsed by the avtpcvfpay, in the same
way done in rtph264pay.

This patch reuses the infra provided by gstavtpbasepayload.c.
2019-07-03 09:59:35 -07:00