Just starting to get into gstreamer so apologies if this is covered in some FAQ. Point me at it and I'll read up offline. I'm about to start developing an audio sink plugin to some hardware. The examples i've found to date appear to be push-based sink plugins. The driver to the hardware I'm interfacing to gets supplied with data via a callback, which would seem to best be supported by a pull based audio sink plugin. So should I have a couple of frames of buffering and go for a push-based plugin, or attempt a pull-based one (presumably with a queue just before the sink?). Pro's and cons please? Thanks in advance. Ian