[AG-TECH] Real Time Music Collaboration

Bob Dixon Bob_Dixon at osu.edu
Wed Mar 13 08:03:18 CST 2002

Thanks for remembering this Marcus. Yes indeed we did the barbershop 
quartet, and it worked wonderfully well. The method used was 
essentially Method 1 described below. We created a CD in advance of 
the event, which contained the songs that were to be sung. We ran 4 
simultaneous point-to-point H.323 video
conferences. The outgoing audio to all sites was the CD audio. the 
outgoing video was the audience. Each singer sang in time and in tune 
with the audio he received. At the concert audience location, all 4 
incoming audio streams were mixed, using adjustable analog delay 
lines to synchronize them. The timings were measured and set in 
advance by having each of the remote sites hold their microphone 
close to their speaker, and playing a metronome out to them all. This
tactic overwhelmed the remote echo cancellers with enough volume to 
make the return ticks audible.
The returning metronome ticks were all displayed on an oscilloscope, 
and then the delay lines were adjusted until all the return ticks 

The return video from each singer was combined in a video mixer, and 
presented as a large projected quad split screen to the audience. The 
singers never saw or heard one another. The conductor (Maestro 
Roberto) was also never seen by the singers, and he in fact just 
conducted in time to the video and audio being displayed to the 
audience. The audience was never aware of any of this, and so it 
appeared to them to be a perfectly normal concert.

The prerecorded CD had a countdown sequence before each song, so the 
singers could start smoothly. Each song track of the CD was started 
manually from the central location as the song announcement was 

The final song was the premiere performance of the Internet2 song. 
This was to the tune of "This land is your land", with new lyrics 
composed by Maestro Roberto. It starts out "This net is your 
net....". It was an audience singalong, and a second large screen 
showed the words to the audience, using karaoke software running in a 

The audience was wildly enthusiastic, clapping and cheering, and then 
getting into rhythmic clapping, shouting "encore, encore!". 
Unfortunately we had not prepared an encore.

We had several redundant systems in operation to record the event. 
Tragically, they all failed. Some of the equipment (including the 
main recording VCR) was damaged during shipment to the hotel. The 
concert hall was not ready when we arrived the previous day to set 
up, despite advance arrangements. So we lost much time waiting for 
power, room audio and network connections, and ran out of time to 
debug everything properly. Also there were only two of us there and
the equipment setup was extensive and complicated. I saw several people in the
audience recording the event with camcorders, but we have never been 
able to find these people and get a copy of their recordings.

I would be interested in following this discussion.


At 9:28 AM +1100 3/13/02, Markus Buchhorn wrote:
>G'day All
>[I've left the whole message in for Bob's (cc'ed) information...]
>During the previous H323 Megaconference(2) event, Bob Dixon (OSU) 
>coordinated a virtual barbershop quartet (or tried to?). I don't 
>know how it was done, or how it worked out as I missed the event, 
>and there's nothing on the web site about it. Bob - some Access Grid 
>folks are looking at a similar project, and would be interested to 
>know what you did and how it went (if it went).
>         Markus
>At 01:07 PM 12/03/2002 +0000, Osland, CD (Chris) wrote:
>>I see, by looking back at an e-mail I sent last week, that I
>>rather buried two possible solutions to the latency problem.
>>May I restate them, as I think they have some mileage...
>>Method 1 - "Audio Conductor"
>>One site produces in effect an augmented click track, with enough
>>of the melody for people not to get lost.  This goes to all
>>players, who just listen to this and play in time with it.
>>Hence the term "audio conductor".  All feeds from the players
>>go to one site, which retimes them individually, having worked
>>out which node is last to arrive.  This site then rebroadcasts
>>the resulting mix to everyone but only the listeners, not the
>>players, listen to it.  This is rather like an orchestra doing
>>the score for a film, each listening to parts of what is
>>happening on headphones.
>>Method 2 - "Round Robin"
>>One site lays down a bed for the piece - maybe a drum track.
>>A second node listens to this and plays (=adds) their player(s)'s
>>contributions.  This feed goes to the third playing node, which
>>does likewise.  The feed from the n'th node (if n nodes are playing)
>>is the one that all the listening nodes listen to.  Typically the
>>n'th node will be the vocals (if the piece is pop/folk/rock/vocal).
>>This is rather similar to the multi-layer approach taken by
>>groups that do the drums in LA, the guitars in London and
>>the vocals in the Seychelles, by lugging a 24-track tape or
>>ProTools system around.
>>It has been pointed out to me over lunch that many classical
>>pieces (e.g. string quartets) cannot be done this way, as at
>>different times each of the instruments may lead the timing;
>>possibly a non-recorded harpsichord continuo would be appropriate
>>in this case!
>>The person who noted this is a classically trained cello player,
>>so I guess I can lob that in as a possible contribution from
>>here as well.
>>Chris Osland                         Office tel: +44 (0) 1235 446565
>>Digital Media and Access Grid      Medialab tel: +44 (0) 1235 446459
>>BIT Department             Access Grid room tel: +44 (0) 1235 445666
>>e-mail:   C.D.Osland at rl.ac.uk               Fax: +44 (0) 1235 445597
>>CLRC Rutherford Appleton Laboratory (Bldg. R18)
>>Chilton, DIDCOT, Oxon OX11 0QX, UK
>Markus Buchhorn, Information Infrastructure Services,   | Ph: +61 2 61258810
>Markus.Buchhorn at anu.edu.au, mail: CompSci,CSIT Bldg #108|Fax: +61 2 61259805
>Australian National University, Canberra 0200, Australia|Mobile: 0417 281429

Robert S. Dixon, Ph D, PE

Chief Research Engineer

Ohio State University, Office of the CIO
Ohio Academic Resources Network (OARNet)

Postal address:                  Office Telephone: 614-292-1638
Office of the CIO                Lab Telephone:    614-292-7425
Ohio State University            Fax:              614-292-7081
1971 Neil Ave                    Email:        Bob_Dixon at osu.edu
Columbus, OH 43210

More information about the ag-tech mailing list