There is no point to that, the code is already factored in such
a way that erase_display_memory is inserted at the correct time,
including while loading the next pop-on captions in non displayed
memory.
Otherwise, it would deadlock when releasing first the primary stream and
then the secondary one, because it will wait in the "secondary stream"
handling for the main stream to be fed, which will never happen.
Locking order of state and settings was inconsistent, and causing
deadlocks. Fix and document it, consistently drop locks before
chaining up events / pushing and avoid sequentially unlocking /
relocking settings in the same local code path.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/539>
The default behavior for the transcriber is to output text buffers
synchronized with the input stream, introducing a configurable
latency.
For use cases where synchronization is not crucial, but latency
is, the lateness property can be used instead of or in combination
with the latency property, in order to introduce a configurable
offset with the input stream.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/534>
This enables a use case for preparing slow to start up sources
ahead of time in a live cueing system, where a stream is scheduled
to start at some point in the future, and the application wants to
make sure it is ready for prime time by that time, instead of
spinning it up at the last moment and waiting for the stream to
actually come up.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/515>
When only the backup pad is receiving buffers, and the primary
pad is a bit slow to start up (eg network source with buffering),
it makes for a better UX to output buffers from the backup pad
while waiting for the network source to make its move.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/515>
This new element puts together some of the elements we've written
in recent times (awstranscriber, tttocea608, textwrap, cccombiner)
into a convenience high-level element.
The design of the element is AV in -> AV (+ CC metas) out.
The element exposes property to set and unset a "passthrough" mode,
during which the transcriber element's state is set to NULL but kept
in the bin, in order for the user to be able to set properties on
sub elements no matter what the current mode is, using the
GstChildProxy interface.
In addition, the element ensures that the latency it reports stays
fixed so that playback continues uninterrupted.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/528>
As awstranscriber might in theory push out gap events without
any flow of input data, it needs to send its mandatory events
(stream-start, caps, segment) independently.
In addition, track a start time and use it to offset the 0-based
timestamps returned by AWS in order to output buffers timestamped
in the running-time domain, and perform item timing adjustment
only when dequeuing, instead of when queuing.
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/525>
and use 0 instead of -1 to disable accumulation. In practice 0
already had the same effect, and this lets us get rid of unsafe
code that actually breaks inspection of the element.
<https://aws.amazon.com/blogs/machine-learning/amazon-transcribe-now-supports-partial-results-stabilization-for-streaming-audio/>
Amazon seem to have realized the previous iteration of their API
made it difficult to identify items from one result to the next,
which made the element much more complicated than it should have
been. With that new "stability" option, we can enqueue items as
soon as they stabilize, and simply rely on the current index in
the transcript to output them exactly once.
This also means the "use_partial_results" is now useless, as there
will be no difference in accuracy between a non-partial result and
and of its stable items that might have been pushed from previous
partial versions of the result.
The property is removed, instead a new option is exposed to let
users control how fast results should stabilize.
This greatly simplifies the code, and also improves the output as
punctuation doesn't need to be randomly discarded anymore.
