Initially, GstAudioDecoder calls @start when the decoder element
is activated, which allows subclass to perform any global setup.
Base class (context) parameters can already be set according to subclass
capabilities (or possibly upon receive more information in subsequent
@set_format).
GstAudioDecoder calls @set_format to inform subclass of the format
of input audio data that it is about to receive.
While unlikely, it might be called more than once, if changing input
parameters require reconfiguration.
GstAudioDecoder calls @stop at end of all processing.
As of configuration stage, and throughout processing, GstAudioDecoder
provides various (context) parameters, e.g. describing the format of
output audio data (valid when output caps have been set) or current parsing state.
Conversely, subclass can and should configure context to inform
base class of its expectation w.r.t. buffer handling.
Base class gathers input data, and optionally allows subclass
to parse this into subsequently manageable (as defined by subclass)
chunks. Such chunks are subsequently referred to as 'frames',
though they may or may not correspond to 1 (or more) audio format frame.
Input frame is provided to subclass' @handle_frame.
If codec processing results in decoded data, subclass should call
@gst_audio_decoder_finish_frame to have decoded data pushed
downstream.
Just prior to actually pushing a buffer downstream,
it is passed to @pre_push. Subclass should either use this callback
to arrange for additional downstream pushing or otherwise ensure such
custom pushing occurs after at least a method call has finished since
setting src pad caps.
During the parsing process GstAudioDecoderClass will handle both
srcpad and sinkpad events. Sink events will be passed to subclass
if @event callback has been provided.
GstAudioDecoder class calls @stop to inform the subclass that data
parsing will be stopped.
Subclass is responsible for providing pad template caps for
source and sink pads. The pads need to be named "sink" and "src". It also
needs to set the fixed caps on srcpad, when the format is ensured. This
is typically when base class calls subclass' @set_format function, though
it might be delayed until calling @gst_audio_decoder_finish_frame.
In summary, above process should have subclass concentrating on
codec data processing while leaving other matters to base class,
such as most notably timestamp handling. While it may exert more control
in this area (see e.g. @pre_push), it is very much not recommended.
In particular, base class will try to arrange for perfect output timestamps
as much as possible while tracking upstream timestamps.
To this end, if deviation between the next ideal expected perfect timestamp
and upstream exceeds #GstAudioDecoder:tolerance, then resync to upstream
occurs (which would happen always if the tolerance mechanism is disabled).
In non-live pipelines, baseclass can also (configurably) arrange for
output buffer aggregation which may help to redue large(r) numbers of
small(er) buffers being pushed and processed downstream. Note that this
feature is only available if the buffer layout is interleaved. For planar
buffers, the decoder implementation is fully responsible for the output
buffer size.
On the other hand, it should be noted that baseclass only provides limited
seeking support (upon explicit subclass request), as full-fledged support
should rather be left to upstream demuxer, parser or alike. This simple
approach caters for seeking and duration reporting using estimated input
bitrates.
Things that subclass need to take care of:
Provide pad templates
Set source pad caps when appropriate
Set user-configurable properties to sane defaults for format and
implementing codec at hand, and convey some subclass capabilities and
expectations in context.
Accept data in @handle_frame and provide encoded results to
@gst_audio_decoder_finish_frame. If it is prepared to perform
PLC, it should also accept NULL data in @handle_frame and provide for
data for indicated duration.
This base class is for audio decoders turning encoded data into raw audio samples.
GstAudioDecoder and subclass should cooperate as follows.
Configuration
As of configuration stage, and throughout processing, GstAudioDecoder provides various (context) parameters, e.g. describing the format of output audio data (valid when output caps have been set) or current parsing state. Conversely, subclass can and should configure context to inform base class of its expectation w.r.t. buffer handling.
Data processing
Shutdown phase
Subclass is responsible for providing pad template caps for source and sink pads. The pads need to be named "sink" and "src". It also needs to set the fixed caps on srcpad, when the format is ensured. This is typically when base class calls subclass' @set_format function, though it might be delayed until calling @gst_audio_decoder_finish_frame.
In summary, above process should have subclass concentrating on codec data processing while leaving other matters to base class, such as most notably timestamp handling. While it may exert more control in this area (see e.g. @pre_push), it is very much not recommended.
In particular, base class will try to arrange for perfect output timestamps as much as possible while tracking upstream timestamps. To this end, if deviation between the next ideal expected perfect timestamp and upstream exceeds #GstAudioDecoder:tolerance, then resync to upstream occurs (which would happen always if the tolerance mechanism is disabled).
In non-live pipelines, baseclass can also (configurably) arrange for output buffer aggregation which may help to redue large(r) numbers of small(er) buffers being pushed and processed downstream. Note that this feature is only available if the buffer layout is interleaved. For planar buffers, the decoder implementation is fully responsible for the output buffer size.
On the other hand, it should be noted that baseclass only provides limited seeking support (upon explicit subclass request), as full-fledged support should rather be left to upstream demuxer, parser or alike. This simple approach caters for seeking and duration reporting using estimated input bitrates.
Things that subclass need to take care of: