Feeding raw audio data into a GES timeline

[Daniel Bailey]

Hi,
I am working on an audio synthesis tool, which allows the user to generate
audio to play over the top of a sequence of videos.
I'm writing this in python.
I am using GES to create a simple video sequence editor.
The application needs to be able to play the synthesised audio in sync with
the video.

I have written a simple proof of concept subclass of GstApp.AppSrc which
can do some simple audio synthesis, and feed it into a GStreamer pipeline.
This works fine for me if I just connect it up to an autoaudiosink without
using any GES stuff.

Is there a way to get this gstreamer element to be part of the GES timeline?

I tried to do this using the following method:

Create a Subclass of GES.SourceClip
Implement do_create_track_element on this subclass
do_create_track_element instantiates a custom subclass of GES.AudioSource
the subclass of AudioSource overrides do_create_source and returns a
pipeline that wraps my custom AppSrc.
call layer.add_clip on a layer in a GES timeline to add one of these custom
clips

This method does not work, I see various assertion errors when I try this.

Is this the right approach? Am I doing roughly the right thing? Or is there
a totally different way to achieve what I want?
If this is roughly the right idea then where am I going wrong?

Thanks.
Dan.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/gstreamer-devel/attachments/20210219/585fc347/attachment.htm>