Bridging a Polycom into the AG

[Justin Binns]

A little more detail on the vrm/drm thing - this is using the stock linux
vrm (the same vrm and UCL Vic used on most AG video machines), with an
additional copy of UCL Vic being run for 'display' purposes (by the drm).  
All are UCL vic (actually, all are the same binary ;-).  I *believe*
what's happening is that the default multicast socket behavior under Linux
is to *not* loopback multicast packets (whereas on Windows, where things
do what you'd expect, the default behavior is to accept your own packets),
so the VRM-spawned Vic video streams aren't seen by the display-Vic.  
When Ti hits 'transmit' on the display vic, he's sending duplicate streams
(the display-vic and the video-vic are streaming), but local Vic code is
doing the duplication to display it to the screen.  I haven't hunted
through the code, so I could be wrong, but I think the above is likely...

Justin

P.S.  Not sure what the fix is - we could probably patch UCL-Vic to
explicitly set the socket to allow loopback of multicast, but I don't know
if that's the right thing....

On Sat, 19 Oct 2002, Ivan R. Judson wrote:

> 
> > While lying down resting, I was thinking that it probably 
> > wouldn't be too hard to bridge a polycom unit into the AG so 
> > that a) Remote polycom users could participate in an AG 
> > meeting without an AG and b) AG users could participate in a 
> > h323 meeting.
> 
> It depends on the strategy you employ, but before I poison you with my
> theories, what'd you come up with? If there's an idea we've missed; I'd
> sure like to hear it :-)
>  
> > On a related note, after wiring the uM2 up to be AG-ified and 
> > getting the linux PIG nodes working smoother, I've found out 
> > several things:
> > 
> > - MatchMakers aren't needed. In order to use unbalanced audio 
> > with the Genter, wire tip or left channel to positive on the 
> > phoenix connector and tie ground to negative and ground on 
> > the pheonix connector. Viola!
> 
> Excellent. One less piece to have confusion about.
> 
> > - It looks like the alsa sound drivers are a valid free 
> > alternative to the 4Front drivers, at least on PIG nodes. I 
> > just installed the alsa drivers on personal-ag which has an 
> > i810 chipset and it's full duplex and seems to work quite 
> > nicely. We need to find an alternative to the 4Front drivers 
> > because a) Their should be an open source alternative re: 
> > latest discussion on Groove and b) the es1370 chipset is no 
> > longer being made and c) most computers are coming with 
> > onboard sound which by and far is the i810 or Via chipsets
> 
> This was mildly confusing to me, but I think the gist of it is: standard
> onboard audio is now sufficient for AG requirements and works with free
> drivers (alsa). 
> 
> Caveat: Are those the default drivers with any (hopefully most) linux
> distributions?
> 
> > - The linux PIG drm doesn't by default transmit video. The 
> > effect this has is that you can't get your own video feed 
> > locally until you check the transmit from the menu. Oddly 
> > though your video is being fed because remote users see it 
> > fine. I guess that's becuase the vrm is transmitting.
> 
> This sounds very bizarre, is this with OM vic or ucl vic? It feels like
> a familiar problem I saw with OM vic once.
> 
> > - A linux port of the genter software would come in *very* 
> > handy on the uM2 AG node.
> 
> Yes, although I don't know how much day to day fiddling with the gentner
> is required. If you have a gain on rat for each site, you can do mixing
> there -- you don't have to fix your audio. In fact, theoretically once
> the gentner is set up we shouldn't have to fiddle with it at all (this
> assumes our mics are not moving and people are basically in the same
> position and orientation every time, which happens to be the case 75-80%
> of the time thanks to furniture).
>  
> Good work!
> 
> --Ivan
> 
>