hi jim,

thought you would get a kick out of the following discussion regarding an upcoming Hub CD-ROM version recording of Variations II, audio and video. we're planning to use the Convolatron at XeroxParc to do auto-spatialization of the audio on the basis of dot positions in Cage's score.

here's more musical banter of the future!

------------------------------------- from Bill T.: [Thibault]

actually, S. Cahill called the hub performance "striking." Solidarity with oakland teachers!

If we want to do var2 in 3d, I can get something happening on this indy, and with probability .9999999 lug it to parc or whatever. Or, if the convolvatron takes midi, we can record the control track into a sequencer. I'd feel better bringing the box incase the sequence turns out not to work right once we're on site.

For 3d its spheres and planes. visually, one question is: what's the point? (of view) It seems to me that the POV of the audio should be the same as that for the video. if we view from the listener's position, and put her in the center of the cuboid, then hubster-spheres behind her wont be visible... but they will be audible. I kind of like that. otherwise, if the viewer is outside the cube, we could see everything, but not really be "Immersed is hublike goo"

but the POV is easy to change once the mechanism is in place. porting the "straight" cage version as 2d would be easier, but again the issue of POV arises. If we see the score, is it as if the listener is looking at a wall and hearing the hubster-dots emit sound? or is the listener on the wall herself? (glued to a wall of hublike goo)

From markg@island.com Thu Nov 30 11:41:19 1995

>or, if the convolvatron takes MIDI, we can record the control track >into a sequencer.

negatory on that good buddy, at least last time i used it. it's a DIY serial protocol... runs at 14.4 or less. client code is fairly portable between unix machines, and i had a version running on my amiga at one point.

>if we view from the listener's position, and put her in the center >of the cuboid, then hubster-spheres behind her wont be visible... >but they will be audible. I kind of like that. >otherwise, if the viewer is outside the cube, we could see everything, >but not really be "Immersed is hublike goo" > >but the POV is easy to change once the mechanism is in place. >porting the "straight" cage version as 2d would be easier, >but again the issue of POV arises. If we see the score, >is it as if the listener is looking at a wall and hearing the >hubster-dots emit sound? or is the listener on the wall herself? >(glued to a wall of hublike goo) >

you could easily (from the convolvotron's end of things) have it both ways... immersion in hub goo plus moving out of hub-goo-area.

one note from a guy who's hacked convolvo-code, make your hub-goo-area really big... as i remember you can define spaces from inches/centimeters big to miles/kilometers big... making it a small area (like 100 meters) produces a less dramatic effect than a large area like a square kilometer. making the convolvotron sound good is not easy... it could take a day or more of tweaking code... and even then, e.g., it might work really well for chris but not for john... depends on how closely one's head matches the standard HRTFs.

of course, you guys probably already know this... i know mr. stone does!

-mark

From phil_stone@maillink.berkeley.edu Thu Nov 30 11:45:16 1995

I think immersion in the middle of the hublike goo is the key here (what a great piece or CD title in there somewhere).

Even though the front/back up/down aspect of the Convolvotron is its weakest effect, it *is* there, and if we don't take advantage of it, then we're just doing a kind of fuzzy panning. Let's shoot the works, and if it's possible for you, Bill, go 3D, and put the point-of-hear (and view, I guess, for the video - if the visual does not corroborate the audio, the audio spatialization effect falls apart) in the center of the piece-space.

This sounds ridiculously hard for the graphics, but being the aurally-biased person I am, it's the sonic effect I'm most excited about here, and don't want to sacrifice that to the needs of the graphics. On the other hand, if you *could* pull off a center-of-sphere point of view, I think it would be beautiful (hey, stereo glasses, anyone?)

As far as I know, the Convolvotron does not accept MIDI, just RS-232, despite my constant suggestions to Scott Foster (he just doesn't see the Convolvotron as a studio effect - it's a VR/lab thing to him).

From tebo@mcs.mcs.csuhayward.edu Thu Nov 30 11:50:33 1995

I'd blotted out my agonizing convulvatronic experience. forgot it was serial (hoping it had been changed). diving headfirst into a kilometer of hubgoo, then pulling out, over and over and over......ahhhhhhh.

From phil_stone@maillink.berkeley.edu Thu Nov 30 12:01:52 1995

>one note from a guy who's hacked convolvo-code, make your hub-goo-area >really big... as i remember you can define spaces from >inches/centimeters >big to miles/kilometers big... making it a small area (like 100 meters) >produces a less dramatic effect than a large area like a square > >kilometer.

Good point. I think the Max patch for converting MIDI to RS232 is a good way to go, because we could easily incorporate a scale parameter to muck with the space size so that it sounds good with the scale of the piece. Too big a space might dwarf our already nicely-thinned realization of Variations II. Two small, and as Mark points out, the spatialization effect is weak.

>making the convolvotron sound good is not easy... it could take a day or >more of tweaking code... and even then, e.g., it might work really well >for chris but not for john... depends on how closely one's head matches >the standard HRTFs.

Also note that you can "learn" the standard HRTF, but this is difficult without visual or some other form of reinforcement. Even without exact localization, I think some spatialization will result, though, esp. in the left/right axis.

Phil

-------------------------------------------------------