On Monday 29 September 2003 12:50, Rob wrote: >On Monday 29 September 2003 12:50, Robert Jonsson wrote: >> Anybody know of an application that allows streaming of midi >> and/or audio over the net for the purpose of allowing several >> people to jam together? > >I seem to remember something like that for Windows too, but >remember that latency that would be more than acceptable for >gaming (30-40ms) could make it impossible to jam as you're >envisioning. Yes, I remember someone (non-technical) telling me about this great system that allowed musicians on both coasts of North America to perform a "live" piece together, across the internet. It was some kind of university project. They even pulled out a newspaper or magazine article about it. I read it through several times. I could not believe that you can get latencies (esp. cross-continent) down low enough to allow "interactive jamming", where both sides hear each other in real time. Let's see (calculating on the back of an envelope): 3000 miles / 186000 miles/second... that's nominally 16msec... that could be tolerable... but what about slowdown due to dielectric (speed of electric fields is less than speed in vacuum)? what about delays in electronic circuits? what about store/forward digital gear? At one point some traffic went via sattelite, which adds 2 x 22K miles (or about 1/4 second). Hmm, that's why the delay on some speech circuits (like when I phone my sister in the Dominican Republic) is very noticeable! Lately, I think ground fibre is cheaper (and faster) than satellite. For the moment, ignoring costs, I'm not even sure the network wizards are able to splice together a dedicated circuit coast-to-coast with audio hi-fi stereo bandwidth, even for "proof of concept". Even if they could (like a permanent phone call?), there would be little point, because in a digital (internet) network there would be real traffic, and hence variability (jitter), which can only be smoothed by buffering and delay. Stutter is usually worse than delay. So, I conclude that I'm mystified! Huh? Now, if one is only concerned with one-way traffic, one can "cheat"! I concluded from the article (and thinking hard about it) that they must have used one site as a "reference site", piped their (partial) performance across the continent (with whatever additional delay/buffering), and then had the other orchestra "dub in" their part, and have that played at the 2nd site for their "live" audience. Or the audience might have been at a 3rd site, at this point it does not matter, just as long as it's not the 1st site! I seem to recall that the audience was seated in an auditorium on 2nd coast. So, yes they were "playing together" in some sense. And the audience was hearing the performance "live". However, I cannot believe that the orchestra at the 1st site was able to hear the 2nd site "live" at the same time they were playing? Has anyone else heard about this? Details? Thoughts? p.s. I used to do some sound recording for 16mm newsreel film stuff, decades ago. Have you (with headphones on) ever tried to speak into a Nagra tape recorder with a true read head after the write tape monitoring head? You hear yourself about 1/2 second later. I had to take my headphones off, so I wouldn't hear a delayed "echo". I think that is also true for "real" (long delay) echo in recording? It can be paralyzing! Is that like stutterers? p.p.s. I have recently been thinking a bit about psycho-acoustics, as I'm (re)learning some guitar playing. If you consider nerve transmission speeds, being able to play those real fast weedle-weedle-weedle guitar leads would seem impossible. What must be happening is that you are telling your fingers to move a fraction of a second before they actually move. Now add in the long echo delay, and I suspect that's too much to handle: 3 time bases: what you want to play, what you are playing (feel?), and what you hear. Comments? -- Juhan Leemet Logicognosis, Inc.
