[Bug 740236] New interleave2 based on GstAggregator and create GstAudioAggregator from audiomixer

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=740236

Sebastian Dröge (slomo) <slomo at coaxion.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEEDINFO                    |NEW

--- Comment #60 from Sebastian Dröge (slomo) <slomo at coaxion.net> ---
(In reply to Olivier Crête from comment #59)
> > @@ +731,3 @@
> > +
> > +  if (!GST_BUFFER_PTS_IS_VALID (inbuf) ||
> > +      aggpad->segment.format != GST_FORMAT_TIME) {
> > 
> > We should require properly timestamps input buffers
> 
> In live pipeline, it's convenient to be able to send out buffers with a
> GST_CLOCK_TIME_NONE pts, to be mixed asap, for thing like sound events and
> stuff.

The meaning being that it perfectly aligns with the samples of the previous
buffers? Sounds reasonable, yes.

But don't support non-TIME segments please :)

> > ::: gst/audiomixer/gstaudioaggregator.h
> > ::: gst/audiomixer/gstaudiomixer.c
> > @@ +393,2 @@
> > +  if (ret)
> > +    gst_audio_aggregator_set_sink_caps (aagg, GST_AUDIO_AGGREGATOR_PAD
> > (pad),
> > 
> > Why? Shouldn't that just be the default behaviour for setcaps()?
> 
> The base class doesn't have a setcaps at all, it just lets the subclass
> handle the caps event.

I see. Maybe that should be changed then :) Looking at videoaggregator, it is
also a bit special there. We should make it sensible in both and also the same
way. But don't block merging this patch on that, let's just get it in before
more changes are added to audiomixer.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.