AW: AW: [Uni-verse] Re: OSGVerse

Thu Apr 6 14:48:44 CEST 2006

Technically, it is for me no difference to update my own object node and
send the transformation command, or to update your UVAS object node and send
the transformation command. I only have to know which one is your
favorite... :)

In my opinion it would be more flexible to update the renderers objectnode
and UVAS will then scan the avatar transformation data of the renderer,
because then you listen to more than one renderer in your system easily. You
can then choose, what renderer you want to follow with your audio
calculation. You would not have the need for a more complex tag definition,
that every renderer would have to understand, but could choose by yourself,
if you want to calculate the audiostream for this or that connected device.
The renderer will save its camera position into the avatar object node
contineously. This information will be sent through Verse - and with that to
your UVAS. There you only would have to listen to the renderers avatar nodes
transformations.
And for you it does not make a big difference to look at your or any other
object node internally, doesn't it?

But maybe there are better reasons that I do not see right now to do it the
other way... ;)

Concerning the orientation:
If I update the rotation of the avatar node this problem is solved, isn't
it?
And that is fortunately not the big challenge...I planned to integrate this
functionality anyway.

Best Regards
Marcus

Marcus Hoffmann
Dipl. Ing. for Mediatechnology
Dep. Realtime Solutions for Simulation and Visual Analytics
Fraunhofer Institut für Graphische Datenverarbeitung
Phone: +49 (0)6151 155 - 639
Mail: marcus.hoffmann at igd.fraunhofer.de

> -----Ursprüngliche Nachricht-----
> Von: Peter Lundén [mailto:plu at tii.se] 
> Gesendet: Donnerstag, 6. April 2006 10:43
> An: Marcus Hoffmann
> Cc: consortium mailing list
> Betreff: Re: AW: [Uni-verse] Re: OSGVerse
> 
> Hmm... think we missunderstood each other. What I refere to 
> as orientation is identical to the information in 
> verse_o_transform_rot_real32.
> 
> It is an intressting discussion about who is responcible for 
> moving the avatars. For me its not a technical issue but a 
> conseptual. There are more than one way to view the problem. 
> My perspective is that the acoustic simulation system is a 
> back end system not realy visible (but audible :-) ) to the 
> user. There is then a front end system, the navigator, which 
> is responcible for the navigation and that system should 
> control the listeners position in the world. The user should 
> be able to manipulation everything from one client (or it 
> should at least be percived as  a single client). If the user 
> is forced to also be aware of UVAS everything gets mush more 
> complicated from the users point of view.
> 
> I think there is some consensus in that clients who are some 
> kind of renderer then its avatar represents the 
> camera/listener. If this is agreed up on then somebody else 
> then the client himself has to drive the avatar if its a 
> passive client. What do the KTH people say about this?
> 
> Then the qustion in our case is who is controling the avatar 
> of UVAS? I still think its the navigator. But technicly its 
> simple to follow any verse object node, given it has its 
> position and rotation updated often enough.
> 
> --PLu
> 
> Marcus Hoffmann wrote:
> 
> >Hi Peter,
> >
> >It will be included until the middle, latest the end of next week. 
> >Sorry for the delay.
> >We will - in the first version - upload the transformation 
> of the our 
> >avatar-object-node using the common verse commands.
> >(verse_o_transform_pos_real32, verse_o_transform_rot_real32).
> >Wouldn't it be a more general way, if you would get the 
> transformations 
> >from that avatar-node in your code (because you have 
> knowledge of our 
> >object node, when we connect)?
> >Or don't you download the other objectnodes? (i think this wouldn't 
> >drop too much performance on your side, object nodes generally are 
> >small verse
> >objects)
> >
> >If this would be not possible, we would have to identify 
> your node on 
> >the renderers side (via the tag) and would have to update its 
> >coordinates, did I get that right?
> >
> >For the orientation:
> >OpenSG has the possibility to give us the data needed to define the 
> >orientation. We just would have to find a way to transmit 
> that to you.
> >Following the 2 approaches I described above for the 
> >transformation-information-submission, we could just 
> establish another 
> >tag for that information, in our object node, or if this 
> version is not 
> >possible in your object node....
> >Then you could get the orientation information everytime you want, 
> >because it's up-to-date.
> >
> >Best Regards
> >Marcus
> >
> >Marcus Hoffmann
> >Dipl. Ing. für Medientechnologie
> >Abt. Echtzeitlösungen für Simulation und Visual Analytics Fraunhofer 
> >Institut für Graphische Datenverarbeitung
> >Phone: +49 (0)6151 155 - 639
> >Mail: marcus.hoffmann at igd.fraunhofer.de
> >
> > 
> >
> >  
> >
> >>-----Ursprüngliche Nachricht-----
> >>Von: uni-verse-bounces at projects.blender.org
> >>[mailto:uni-verse-bounces at projects.blender.org] Im Auftrag 
> von Peter 
> >>Lundén
> >>Gesendet: Mittwoch, 5. April 2006 15:40
> >>An: Marcus Hoffmann
> >>Cc: consortium mailing list
> >>Betreff: [Uni-verse] Re: OSGVerse
> >>
> >>Hi,
> >>
> >>What is the status of the OSGVerse renderer. Can it handle the 
> >>navigation now?
> >>
> >>As we talked about, it must also be able to attach the UVAS 
> listener 
> >>object. As discussed in Stockholm the UVAS listener will be named 
> >>'uvas'
> >>and will have tag 'listener' in the 'uvas' tag group ( in 
> the future 
> >>we have to find a better naming scheme that works with more 
> then one 
> >>listener). You should find that object and update its position and 
> >>orientation according to the camera in the visual renderer.
> >>
> >>Have you sorted out the problem of how to get the direction? 
> >>I dont think there is a reference direction specified for Verse but 
> >>UVAS ceartenly have one, the reference direction vector is (0 0 -1)
> >>
> >>We need to be able to test the whole demo system with all the 
> >>components before the review and is just a few weeks left.
> >>
> >>Best regards,
> >>--PLu
> >>
> >>
> >>    
> >>
> >
> >_______________________________________________
> >Uni-verse mailing list
> >Uni-verse at projects.blender.org
> >http://projects.blender.org/mailman/listinfo/uni-verse
> >
> >  
> >
> 
> 