From: jeff@is.s.u-tokyo.ac.jp (Jeff McAffer) Subject: Re: SCI: Sound in Virtual Worlds - MIDI Date: Mon, 5 Jul 1993 16:59:21 GMT Organization: University of Tokyo / Object Technology International In article <218igu$25n@news.u.washington.edu> mpelzshe@man104-1.UCSD.EDU writes: >My experience so far with newsgroups such as this one has been that >when you leave the implementation part out, people lose interest >rather quickly. Not me. Rather, I get bored when people postulate implementations before they have even figured out a design. It's ok to *do* it that way for yourself, just gets sorta boring to talk about cause you are always dealing with the latest implementation problem and losing sight of the real problems (like who cares if the problem feature is even there?). >delay after the command is executed. The very definition of a voice, >furthermore, is too ambiguous without specific reference to the signal >processing resources the voice has at its disposal. Is the voice >controlling a digital oscillator? A sample playback system? A MIDI >synthesizer? Perhaps it would be better to specify "instruments" and >"performers", rather than the single term "voice". In other words, >defining an API in a vacuum is a waste of time, IMHO. True about the vacuum part. However, done properly (tm), how a voice (or whatever) is represented (implemented) is of little interest. What is interesting is what voices can do and what others can do to them. Should they even exist? In what situations? That's the API part. You mention the NeXT sound stuff I agree that its nice. One of the things that we should get from that is that its object-oriented. Conceptually it doesn't matter if some "sound" is a MIDI file, a waveform, samples... You can tell it to play here or there, ask it for what it would output (the outgoing samples), ... >In response to jeff@is.s.u-tokyo.ac.jp (Jeff McAffer), who wrote: > >> Adopting MIDI as a protocol is pointless. As I understand it MIDI >> is quite low level. It talks about notes and instruments and >> attributes ... It is a remote device control protocol. It is a >> MUSIC protocol, not a sound protocol. > >Unless you're talking about the General MIDI spec, the MIDI spec knows >nothing about "instruments and attributes". It is, as you say, a >remote device control protocol, nothing more, nothing less. I think But of course. Its a remote device control protocol aimed at controlling remote musical instruments (*MI*DI). It has nothing to do with sound. Recognizing this, I made my original statement which was essentially asking, "why is MIDI even being mentioned?". >you're real complaint is that MIDI isn't low level *enough*! (But No, it's that we are even talking about it. >then, I tend to think of Music as hierarchically "higher" than Sound.) Not all sound is music but all music is sound (using conventional definitions). Put it another way, without sound there is no music (except written...). Sound is more general and thus higher-level. Music is *one* organizations of sound. The high pitched squacks your modem/fax makes is another. Which is more organized? does it really matter in this context? The point here is that if we implement general sound in our worlds, then by default we can have music, the opposite is not true. Sound is a prerequisite for music, not a subclass. >> IMHO, music has nothing to do with computer >> generated realities except that we may want to have some music >> integrated into our worlds. (sure there are some music-specific >> applications but that is not the driving force here). > >Assuming computer-generated realities are a subset of Artificial >Realities, (i.e., film, theater, TV, etc.) I beg to differ with this >assertion. Music in the broad, 20th century sense (as "organized >sound") is an extremely important carrier of information. A world >without music sounds really dull to me! You missed my point. First, computer-generated realities is my lame attempt at a term which does not offend my feeling of experience when I encounter world labelled as such. That is, if I come out of a session in a HMD... and am faced with the label "Artifical" that bugs me. It was damn real to me, just computer generated. But I'm anal retentive about names. I was not trying to say that sound was unimportant or undesireable or useless. In fact, its role as an input to humans is one of the major directions of my current research. All I was trying to say is that sound data is not essentially different in this context than any other kind of data. Many of the issues in "Sound in VR" are really issues of "data in VR". Solve the general problem in a general way and you get a solution (or a good part of it) to the more specific problems. So we would do well to think about what issues relating to sound we want to deal with (i.e., those which are not going to be part of a general solution) and then set out to solve them. -- ato de, |m -- A strange and bizarre beast lurks under many a bed. Under mine there is a can of Pringles!