[Bug 797169] compositor: Remove crossfade property and implement a more extensible interface like in glvideomixer

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=797169

--- Comment #28 from Sebastian Dröge (slomo) <slomo at coaxion.net> ---
(In reply to Thibault Saunier from comment #26)
> (In reply to Sebastian Dröge (slomo) from comment #25)
> > (In reply to Thibault Saunier from comment #23)
> > 
> > > I do not think we want this kind of restriction at all, the end user can
> > > transition whatever he want, no reason to not let it do that!
> > 
> > Not having such restrictions means more complicated internals and it
> > restricts how you can change the internals / extend the internals :) But
> > that's for somewhere else to discuss I guess.
> 
> Well, this issue is exactly about that... you want to simplify the
> internals, at the cost of flexibility and even removing useful features (2
> pads only (which is arguably not a mandatory feature), and now adding
> restriction which must *not* exist from a end user video editing app
> perspective).

My point here about compositor is not that. Half of the story is the code
complexity, the other part is a non-intuitive interface that you wouldn't
expect from something called a "compositor".

We generally build elements that are simple and do one thing, and then you can
combine them with others or use them multiple times even. And by doing that you
can still get whatever blending behaviour you want without making the API
weird.


For cross-fading multiple pads might make sense, thinking of something like a
crossfader+concat for example but that wouldn't be useful for pitivi. I don't
know what pitivi exactly needs for cross-fading but if there's a use-case for
cross-fading the first two pads, and then the result and a third pad, then
sure, go for it. But that would also be implementable by two 2-pad crossfade
elements then and would perform exactly the same operations then (and be more
or less the same performance-wise).

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