[Bug 766236] rtp j2k payload/depayload messes up colours in sample pattern

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=766236

--- Comment #137 from Sebastian Dröge (slomo) <slomo at coaxion.net> ---
(In reply to boxerab at gmail.com from comment #132)
> (In reply to Sebastian Dröge (slomo) from comment #126)
> > The parser should probably also require either the sampling field, or the
> > colorspace field. It can reconstruct the other one from the existing one :)
> 
> How do we specify that the sink pad requires one or the other of two fields
> ? I thought that if a field is specified in the static caps, then it must be
> present upstream ?

You would put two structures in there, e.g.
  image/x-jpc,colorspace=(string){...},
  image/x-jpc,sampling=(string){...}

The caps with either of the two (or both) are compatible.

(In reply to boxerab at gmail.com from comment #134)

> > And should probably also make use of the sampling field in the decoder... to
> > distinguish between BGR and RGB for example.
> 
> Btw, a JP2 stream can be parsed to determine the channel order.

Yeah, but for others we have no idea :) Our colorspace field does not give
enough information to properly decode all kinds of images, while the sampling
field does. For 2.0 we should standardize on that one only, but for backwards
compatibility reason we need to handle both (and in case of only colorspace
fall back to guessing something).

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