[Bug 745107] glimagesink: implement GstVideoOverlayCompositionMeta

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=745107

--- Comment #84 from Jan Schmidt <thaytan at noraisin.net> ---
(In reply to Nicolas Dufresne (stormer) from comment #81)
> (In reply to Jan Schmidt from comment #79)
> > When rendering out to side-by-side or top-bottom frame packings, if we're
> > blending subpictures, it'll need to be done twice - once for each eye view
> > in the output video. Other frame packings (line-by-line,columns, etc) are
> > probably worse.
> 
> Do you have pipeline and streams, or do you mean you will implement this ?

I'll look for or create a usable stereo stream with a suitable subs track for
testing. The only one I have isn't share-able.(In reply to Nicolas Dufresne
(stormer) from comment #82)

> With the latest patch (and the latest basetextoverlay patch, see bug
> 751157), it's doing the right thing for the non-3D videos. There is a bug
> that initial negotiation does not see the overlay meta suppport initially.
> I'll work on that tomorrow.
> 
> For 3D support, note that it cannot be done if some combiner is in place.
> This should already negotiate the textoverlay to render directly. In
> general, I would not offer the meta if the caps are 3D, so we don't get
> rushed in figure-out what would be the right thing to do.

Yep - at the moment it gets it horribly wrong there. It'll render over the top
of a frame-packed video frame, and then when it gets converted you end up with
half the subtitle showing in the wrong place to each eye.

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