Representation of data processing results as camera devices?

[Nicolas Dufresne]

Le lundi 15 février 2021 à 09:48 +0100, Markus Elfring a écrit :
> Hello,
> 
> Are camera objects available which represent results from data processing
> pipelines
> according to audio and video so that they are not limited to concrete picture
> and sound sensor devices?
> Can sensor devices be constructed on demand?

You question is a little too vague, perhaps due to the very specialized
background. GStreamer framework can be used for any type of stream, even though
some well known streams like video, audio and subtitles have a lot more helpers
and libraries around it.

Even when split, all streams are synchronized though a strict timstamp (and
segment) mechanism. To for metadata, all option are valid. You can model them as
streams, at in-band event (like tags), or attached to audio/video frames (with
GstMeta). It all depends on their nature and their mutability.

> 
> Regards,
> Markus
> _______________________________________________
> gstreamer-devel mailing list
> gstreamer-devel at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel