In the case an aggregator is created and pads are requested but only
linked later, we end up never updating the upstream latency.
This was because latency queries on pads that are not linked succeed,
so we never did a new query once a live source has been linked, so the
thread was never started.
https://bugzilla.gnome.org/show_bug.cgi?id=757548
When caps changes while streaming, the new caps was getting processed
immediately in videoaggregator, but the next buffer in the queue that
corresponds to this new caps was not necessarily being used immediately,
which resulted sometimes in using an old buffer with new caps. Of course
there used to be a separate buffer_vinfo for mapping the buffer with its
own caps, but in compositor the GstVideoConverter was still using wrong
info and resulted in invalid reads and corrupt output.
This approach here is more safe. We delay using the new caps
until we actually select the next buffer in the queue for use.
This way we also eliminate the need for buffer_vinfo, since the
pad->info is always in sync with the format of the selected buffer.
https://bugzilla.gnome.org/show_bug.cgi?id=780682
Sending an event can accepted event if the caps were rejected
because the event could be queued and processed later.
Also send a drain query in the caps test to make sure that the
event has been processed.
https://bugzilla.gnome.org/show_bug.cgi?id=781673
The function needs to be unlocked if any data is received, but only
end the first buffer processing on an actual buffer, synchronized events
don't matter on the first buffer processing.
https://bugzilla.gnome.org/show_bug.cgi?id=781673
With the macOS/iOS implementations, the active thread can change
multiple times over the life of a pipeline which would expose a race in
the thread tracking.
Fix by taking a ref on the active thread while the context is active.
https://bugzilla.gnome.org/show_bug.cgi?id=779202
This ensures smoother playback. It looks weird if we first do a big
jump, then play a couple of consecutive frames, just to again skip ahead
quite a bit because we ran late again.
Far enough here means more than 500ms or 4 times the average keyframe
download time. There is no need to jump ahead by one average keyframe
download time in this case.
This makes playback smooth if the network is fast enough.
When dealing with key-unit trick mode downloads, the goal is to
provide the best "Quality of Experience". This is achieved by:
1) maximizing the number of frames displayed per second
2) avoiding "stalling" as much as possible (i.e. not downloading and
decoding frames fast enough)
This implementation achives this by:
1) Knowing very precisely the current keyframe being download (i.e
more accurate than at the fragment level which might contain more
than one keyfram). This is the new "actual_position" variable
introduced by this commit
2) Knowing the position of downstream (provided by QoS and stored
in the adaptivedemuxstream qos_earliest_time variable)
3) Knowing how long it takes to request and fully download a keyframe
(the average_download_time variable)
Taking those 3 variables into account, whenever a keyframe has been
pushed downstream we calculate a "target time" (target_time variable)
which is the ideal next keyframe time to request so that:
1) It will be requested/downloaded/demuxed/decoded in time to be
displayed without being too late
2) It will not be too far ahead that it would cause too few frames
per second to be displayed.
How far ahead we will request is inversily proportional to how close
the actual position (actual_position) is from the downstream
position (qos_earliest_time). The more is buffered between the source
and the sink, the "closer" the target time will be, and therefore
the more frames per seconds will be displayed (up to the limit
of keyframes_per_second * absolute_rate).
