[LAD] LV2 port groups

[David Robillard]

On Sun, 2009-08-16 at 19:22 +0200, Jörn Nettingsmeier wrote:
> David Robillard wrote:
> > On Sun, 2009-08-16 at 18:27 +0200, Jörn Nettingsmeier wrote:
> >> David Robillard wrote:
> >>> On Mon, 2009-08-10 at 11:02 -0400, David Robillard wrote:
> >>>> http://lv2plug.in/ns/dev/port-groups.lv2/port-groups.html
> >>> P.S. Comments on this from anyone with experience in multi-channel stuff
> >>> would be appreciated if there's any problems.
> >>>
> >>> It's being used to support proper multi-channel streams for LV2 in
> >>> GStreamer (work in progress), in keeping with the theme of getting along
> >>> better with them desktop guys :)
> >>>
> >>> Personally I would like to see this sort of metadata get exposed at the
> >>> JACK level quite a bit (e.g. so things can do auto-connect logically).
> >>> All we'd need is a key/value system...
> >> is anybody using this yet? if not, can i suggest that the ambisonic
> >> groups be renamed to #H#P, and that #H#V be reserved to that new scheme
> >> introduced by travis, as explained in another mail?
> > 
> > (The "re-use" of #H#V there seems crazy to me... (re: your other email))
> > 
> > Anyway, no, nobody is using this yet.  I'm really just defining things
> > at this point so future stuff that needs it can use it easily and get
> > along.  First thing's first and all that.
> > 
> > I will name the groups (multi-channel) H#P.  I suppose I should use e.g.
> > acn1Channel rather than wChannel for the roles (single channel) too...
> 
> yes. with all that rdf magic, is it possible to include a pointer to
> http://ambisonics.ch/standards/channels/ directly in the spec?

Should the order of channels within a group (which mustn't have gaps)
just match the ACN order?  This is different from FuMa.

e.g. directly substituting names gets you something like:

AmbisonicBH1P1Group
"Ambisonic B stream of horizontal order 1 and peripheral order 1."
0: ACN0 (formerly w)
1: ACN3 (formerly x)
2: ACN1 (formerly y)
3: ACN2 (formerly z)

Should this be:
0: ACN0
1: ACN1
2: ACN2
3: ACN3

(No gaps in this case, but H2P0 has channels(*) 0 3 1 8 4, so I'm
asking...)

Thanks,

-dr

(* Unless something's been nuked in the munging)