Conference napalm::commusic_v1

    My guesses:
    
    - Some nut will construct a full size orchestra out of synths
    and samplers, and sequence a complete symphony. I'm not just
    talking about one timber on a synth dedtcated to the string section,
    but a separate voice for each of the 16 violinists. Each member
    of the orchestra would be represented by its own voice. That person
    will debut the 'orchestra' at a college, and will take it on tour,
    and make a fortune like Dave Brubeck. That person will also start
    the sequencer, and 'conduct' the assembled synths live. This guess
    is not intended as a joke, I really think someone will try this.
    
    - The Kurzweil MIDI Harp contoller will debut as the 1989 NAMM show
      for $5000.
    
    - Two months later, Casio will announce a $195 version of the
      MIDI Harp controller.
    
    - NED will seccumb to pressure to produce a more affordable instrument.
      They will announce their Syn-1000 rack mount. It will do everything
      the full price Synclavier will do, except that it will be monophonic,
      and will only sample at 10kHz.
    
    (the following one was lifted from a LERDS-BIM discussion)
    - Harley Davidson will enter the synthesizer market, and will introduce
      their first synthesizer. It will be made out of tubes, and will
      be difficult to start. (like a Hammond B3)
    
    - IBM will enter the synthesizer market. But, true to their style,
      they will attempt to impose a new synth interconnect standard
      SNA-FU (Synthesizer Network Architecture - For Us). The industry
      will panic, and will dump MIDI to conform to SNA-FU.
    
                                                          

    We touched briefly on this in MUSIC I think.
    
    I think that the next wave will be software.
    
    OK, I admit I'm a SW-type and perhaps that colors my vision, but
    I really think that the next wave will be something like a Music
    Workstation.  I think as computers get faster and memory
    gets cheaper, it will be cheaper to implement things in software
    rather than in hardware.
    
    I envision a music workstation where all these things we talk about
    like samplers, synths, effects, mixers, recorders, patchbays, etc.
    will all come on a floppy instead of inside a box.
    
    You can already see the first stages of this.  
    
    	db

    >- The Kurzweil MIDI Harp contoller will debut as the 1989 NAMM show
    >  for $5000.
    
    Don't laugh.  My neighbor's 12 year old son got a MIDI accordion
    for Xmas last year (all I got was a friggin harmonica).  He already
    has outgrown his FB-01 and wants to hook into my network.
    
    	db

    Warfare will break out between the die-hard Harley synth owners
    ("I got a FX-88 with dual carbs and 16 LFOs") and the "sniveling
    rice-algorythm burners". No-one will be safe as leather clad
    gangs of synthesists attempt to put their rivals out of phase.
    
    Edd (who_recently_became_very_nervous_walking_out_of_a_store
         with_YAMAHA_proudly_displayed_on_his_chest. The_large
         person_with_Harley_Davidson_displayed_on_HIS_chest,_despite
         his_not_wearing_a_shirt,_did_not_seem_interested_in_learning
         about_John_Chownings'_discovery.)
                                          

    The next big step can be looked at an two ways, either as a great
    explosion of self contained musicians, or the gradual decline of
    certain instruments. With many MIDI devices, you can get away from
    requiring certain instruments. Horn Sections, String Sections,
    Bass Guitarists, Drummers, Background Singers, and practically any
    other instrument than can be effectively synthesized, will become
    less important, as an emulation can be easily added. These replacements
    may not be as versitile, or may not be as flexabile as the real
    thing, but they will become more cost effective. Since these
    individuals will have less chance to play in studios, or Live
    situations, software will need to exist to emulate them. Some of
    this already exists, but the sheer effort required to build the
    volumes of necessary libraries will be too much. AI will become
    more involved, and styles of music will be emulated, as will the
    relationships that accompany them. If you want to dial a Van Halen
    solo that fits a song, you will be able to get one (maybe not as
    inspired as what you hear from a live person, but good none-the-less!).
    Sequencer software will become more responsive to live situations,
    enabling many semi-random paths to be available, at any given time,
    and musicians will create parts of the songs & let the AI software
    do it's thing. Tweeking odds & ends until it feels right. The 
    average musician will output directly to a CD or DAT tape & then
    add thier Live parts (voices and/or individual instruments) & then,
    decide the next step. This will all be portable, so now what sounds
    good in a studio, will sound the same Live. Musicians will build
    in things that make thier Live music sound unique, and still use
    the same programs. Bands will shrink in size to the minimum quantity
    needed.
    
    Why do I believe this? I guess programming LISP makes me think this
    way, and attempting to do the above adds to the effort. How far
    away are we? Maybe 5 years, maybe 10. But whatever, not long.
    
    							Jens

    The next phase will  be music-loving systems.  these systems will
    listen to lots of CDs, learn music independently from books,
    write music, play the music, record the music automatically,
    distribute it, and then review their own music.
    Humans will be left completely out of the loop, so there'll be
    lots of time to watch TV.
    I remember back when I was interested in music, these questions
    used to be on my mind, but now it's all boring as hek.
    Tom

    
    I gotta go with db on this one.  The music workstation is not just
    a concept,  just the MAC software alone can  be  quite  astounding,
    and for about $500,000 you can get a top of  the  line  Synclavier
    complete with 19" monitor, sequencing,  sampling,  and a  sampling
    rate that exceeds the speed limit at FermiLab.
    
    Coupled with this I am quite certain will be a new  MIDI  standard
    to provide not only a greater bandwidth,  but greater control over
    a wider variety of different midi instruments.
    
    In addition,  look for someone like Wirth to come along and  intro-
    duce a new standard of structured music  for  sequencing  complete
    with looping constructs and the absence of goto directives.
    
    _sjz.

    re: workstations
    
    I think that a good addition to everyones mailing list would be
    Recording Engineer and Producer.  It goes for $24 a year, and it
    covers all kinds of professional music stuff.
    
    The March 88 issue is dedicated to Digital Audio Workstations.
    The following systems are reviewed/compared:
    Audio+Design Recording............Sound Maestro
    Advanced Music Systems............AudioFile
    Digital Audio Research............Soundstation II
    Fairlight.........................III
    For-A.............................Sirius 100
    Hybrid Arts.......................ADAP
    Lexicon...........................Opus
    New England Digital...............Synclavier
    New England Digital...............Direct to Disk System
    Polyphonic........................FX Optical Transfer Station
    Waveframe.........................Audioframe
    
    After reading about all this stuff, I was astounded at what can
    be done *now*.  And most everything is software-based, so there
    are no inherent limitations.  For instance: 3 hours of 50khz sampling,
    and so on.....
    
    Where do I get one?
    
    Ashley

    But all the _good_ music has GOTOs!
    

    re .9: All the _really good_ music is self-modifying.
    
    

    re .10: I once heard music defined as the mental interaction between
    
    	1) What you have heard (of this piece so far)
    	2) What you are hearing
    	3) What you expect to hear
    	4) goto 1
    
    	There should probably be a line for "What you expected to hear"
                   
    
    Given that preconception and experience control perception, this
    model partly explains why musical literacy is non-trivial even for
    listeners. 
    
    	Eirikur

    re:.-1
    
    	The nut's name is James Romeo, and he sequenced a Debussy(?) opera
    on 2 Kurzweil K250s for the Lowell Lyric Opera.
    
    	- Rick
    

    And there is also Larry Fast who programmed up his last Synergy
    album (MC500's?) and then dismantled his MIDI network, trucked it
    to the recording studio, reassembled it, connected everything to
    the mixer, and pressed the go button.
    
    	Eirikur
    

    
>   1) What you have heard (of this piece so far)
>   2) What you are hearing
>   3) What you expect to hear
>   4) goto 1
    
    which could be modified to something like
    
    loop while_I_like_what_i'm_hearing
    	what am I expecting
        what am I hearing
        did I like this
    endloop
    
    _sjz.

    
    And what about mandlebrot music ?  Take a theme and run it through
    a sound fractal,  and run that through a fractal until it dies  in
    an ocean of white noise.
    
    _sjz

    Coming soon:
    
  o Resynthesis:  How many sine waves do you need to reproduce any
    sound with adequate fidelity ?  This technique fits the current
    cheap-processor, expensive-memory climate much better than sampling
    does.
    
  o Music/Production Workstations: Sampler, Sequencer, and Smpte cue-sheet
    like facilities in a box.
    
  o An all-digital-domain porta-studio: i.e. a DMP-7 + DAT or hard-disk 
    digital multi-track recorder.  This could even be cheap if someone
    would get their head out of... Well, you know what I mean.  Akai
    did, but they're chicken.
    
  o MIDI-LANs:  Build a token-ring or CSMA network with "Midi
    Transceivers".  The network can be any speed you like, or could
    even use an existing standard.  The transciever interfaces would
    be standard Midi, preserving all current equipment investments 100%.
    
    
    

    re -.?
    
    I had heard of the guy with the two K250s and the Lowell Opera Co.
    He's close to what I was suggesting, but not quite.
    
    dp

    Well, I do agree that the current trend is towards the audio
    workstation, but I think that this is unfortunate, and is bound
    to change music in ways that one wouldn't necessarily expect.
    
    What workstations, and AI filters, and so forth, seem to do to music
    is widen the noise margins on the input.  In other words, your input
    doesn't have to be as precisely controlled to get acceptable output.
    
    There seem to be two things that can happen if the ``grunge'' of
    music is eliminated.
    
    One possibility is that, the arena of human interest will shift
    elsewhere.
    
    On pre-step time instruments, an artist has to be very controlled to
    produce something which is musical.  One of the most facinating aspects
    of music is just that.  Like professional sports, it is fun to see or
    hear somebody who has the rare gift required to saw/blow/hammer away on
    an inanimate object to produce a soul moving result.
    
    If anybody can do that, then music performance equipment might become a
    consumer industry.  There is no point in listening to any one elses
    music if you can create the same thing yourself.  Audio workstations
    could sell like cars, if music is a fundamental need, or they could
    disappear, if the fact that there is no challenge to the task makes it
    uninteresting.
    
    The other thing that could happen is that a new form of artist could
    emerge, a meta-artist.  The meta-artist would be similar to a modern
    composer, but many of the difficulties of expressing a composition
    would be removed.  I assume that the meta-artist would be able to
    impose a higher order on the ``boring'' components which are availible
    to all people, with the advent of the ideal music workstation. 
    
    I have seen some computer programs which claim to provide an interface
    which is conducive to producing this meta-music--some of the more
    bizzare ``sequencer'' programs that you read about...
    
    I'm not sure that people are ready for the meta-artist.  I suspect
    that the consumerization is likely to happen, and everybody won't
    pay 25 dollars for a concert ticket any more.
    
    Or neither of the above will happen, which is my hope.  I would
    like to see a new generation of raw, difficult-to-control, instruments.
    
    Steph

    
    
    Just to show what can be done !
    
    Amstrad announced this week the home recording system. Which is
    basically a rackmount lo-fi with 2 cassette decks, speakers, a tuner and
    turntable and (here's the killer) a 4 channel mixer which inputs
    to one of the cassette decks which will record and mix 4 tracks
    ala portastudio's. All for 299 pounds, Casio eat your heart out.
    
    				Paul.                                         
                                                                

    
    ..and I think it also has built-in
    REVERB-berb-benb-bonb-boinb-boiinnngggg
    
    Richard.

    My predictions:
    
    o DAT or CD generation at home.
    
    o A *good* spectrum analyzer interfaced to a PC to keep a record
      of how the final mix is.  It could be used to find hot spots where
      certain frequencies are peaking or places where the frequencies
      are particularly heavy on one band.
    
    o The advent of more small-time studios willing to publish stuff
      generated by small-time MIDI enthusiasts with said DAT or CD in
      hand.  In essence, they will find that there is fast money to be 
      made through promoting music that a home studio has been working on 
      for years for free and which doesn't require much investment on
      their part to produce.
    
    o More emphasis on ease of manipulation of sampled data and output
      filtering.  Less emphasis on bit resolution and bandwidth above
      12-bit and 15kHz, respectively.  Mostly I make this assessment
      because things sound good to the ear with these constraints and
      going to 16-bit and higher bandwidth doesn't pack a lot of bang
      for the buck.
    
    o More emphasis on other effects gear.  
    
    o Less distinction between sampling, LA, FM and PD as separate
      technologies.
    
    o More expensive boxes (higher profit margin), but less devaluation
      of synth gear.  This will be due to the manufacturing learning curve 
      on synths, rising production costs and the lack of any radically
      new way to make sound.
    
    o A disappointing lack of support for the fiber optic standard.
      MIDI works and there is plenty invested in it.  And, above certain
      bandwidths the ear just can't hear much difference, especially
      if there is very little noise or distortion due to good filtering
      on output stages or whatever.
    
    o More keyboards that offer 'everything in a box'.  (For somebody
      like me, this means that I will always be looking for the cut-down,
      rack-mount little brother of the new, bit synths that come out.)
    
    o Good samples available by the bucket.  However, the same will
      not hold for good patches.  Instead, there will probably be more
      and better ways for generating good patches for different synths
      given good samples as a start.
    
    Steve

    re: .18
    
    I really don't think that talent, experience and work will cease to
    become a factor in the production of music.   These machines are still
    musician's tools rather than "artificial musicians".
    
    The talents will change however.  Obviously, you won't need to spend
    years in a woodshed with a guitar to produce music with playing as fast
    as Yngwie Malmsteen.
    
    I'd like to think that there'll always be appreciation for someone who
    has developed a skill and repore with an instrument, but I'm worried
    that I'm just being a traditionalist.  Music is definitely changing
    and to deny that is to ignore reality.  It's happening.  We are seeing
    it.  It's just a question of what direction, how far and how long.
    
    	db

    
>  A disappointing lack of support for the fiber optic standard.      MIDI
>  works and there is plenty invested in it.  And, above certain bandwidths
>  the ear just can't hear much difference, especially     if there is very
>  little noise or distortion due to good filtering  on output stages or
>  whatever.                                     

   Bandwidth in the perspective that you are speaking does have its practi-
   cal limitation (i.e. is the human ear capable of detecting  the  differ-
   ence between a 33K sample and a 50K sample).
   
   On the other hand,  almost all current instruments have  analog  output.
   What happens when the sound actually comes out digital ?
   
   Also how many things can I do at once is another story.  You don't  have
   to do a whole lot to saturate a MIDI network,  especially when you start
   throwing controller information out on the line.
   
   Fiber optics may be somewhat of an overkill,  but I  could  easily  see
   some type coax solution.
   
   _sjz.

    
    I've been thinking about how to say this, but haven't come up with
    any articulate way to do so.
    
    I sincerely believe that while technology that makes making music
    "easier" may tend to "cheapen" commercial music even more than it
    is now, there will always be people like me who buy a record or
    pay to see a show so that the artist can communicate his/her art
    to me.
    
    I personally think that the real lo-end consumer-oriented computer
    music gear is going to produce an entirely different effect from
    what you're afraid of.  People who buy this stuff will be using
    it the same way my mother uses her oil paints.  She isn't interested
    in selling her art, and she certainly isn't going to sit at home
    and look at her work instead of going to an exhibition or something.
    
    I think that once this kind of gear becomes less "neat" and more
    in the mainstream, the public will treat it just like mom's oil
    paints, or dad's wood shop, etc.  
    
    It even may have the opposite effect, that of aquainting the public
    with just how easy it is to round up some neat sounds, and punch
    in a sequence.  Then they will know firsthand what "art" is, as
    opposed to what they're doing.  They will *demand* real artists
    performing and recording real art.
    
    Hopefully,
    						-Jeff
    
    

    Re: .24
    
    I agree that enhanced appreciation of the musical art could be an
    effect of more people becoming musically literate.  However, I think
    that it is a ``short-term'' effect of the technology that we are
    envisioning (I'm not holding my breath for it, since it does seem
    very far away).
    
    Your mother's oil paints do not supply any expertise in painting. She
    can attempt to attack the art the same way that anyone else can.  And
    in most arts, producing a master work is the result of inspiration and
    slaving over the raw materials.  People who attempt art and lack either
    or both of these factors in their attempts gain a better understanding
    of the art. 
    
    But, if you add the type of AI mentioned above (the ``give me a
    VanHalenesque solo'' type) then the user will no longer have to be
    inspired, or slave over the raw materials of yore, which are, in music,
    notes a phrases.  There will be new raw materials perhaps, such as the
    format and production. 
    
    Personally, I most enjoy the individual notes and phrases in music,
    rather than the overall concept.  I am a short-sighted listener in this
    respect, but I rarely listen for the form of a piece (I can, it just
    doesn't usually interest me), so I have some reason to believe that
    making this particular aspect of the art trivial will either lessen my
    interest in music, or lessen my interest in other people's music. 
    
    Steph

    
    	My guess is the universal digital signal processor, capable
    of generating *any* sound, or group of sounds in real time. A
    recording which reproduces a recorded sound will be old hat and
    looked at as a waste of valuable storage space. Instead, the
    artists recording will be synthesized in real time by the "record
    player" itself...in stereo, of course.
    	Since the player will be dealing only with a bunch of instructional
    descriptors, instead of each and every sample of the real music,
    much more music will "fit" within the given media. A CD might last
    for an astronomical amount of time, cause of the inherant data
    compression using the technique.
    	You wont be able to steal by sampling someone's "sound", you'll
    only be able to steal the instructions which in turn generate that
    sound. This DSP synth will be universal and immaculate - any previous
    format can be successfully converted to it's required instructional
    codes. You specify only the detail level needed; a voice would need
    lots of detail, a fuzzed out guitar, perhaps very little. And oh
    the "glossing" and smoothing you can do, just with that function...
    
    	Joe Jas 
    

    
    ....and Fesken's Music Inc. releases "1001 patterns for
    percussive performance"....
    
    on an Aids-protected disk.
    
    smile.
    

    A coupla random observations:
    
    re .18 - as db has already noted, powerful tools do not automatically
    confer skills on their users.  Having a modern word processor with
    spell checking and grammar checking and the like does not make you
    a writer.  In fact, I believe the quality of much writing has *gone
    down* because "writers" now rely on the tools to do their "thinking"
    for them.  Spelling checkers don't know the difference between "there",
    "their" and "they're", so we now see these words misused even though
    they're spelled right.  A real artist can take extraordinary advantage
    of powerful tools - there's no virtue in "slaving over the raw
    materials" (the old "A for effort" canard).  A lot of effort doesn't
    imply excellent results any more than a little effort implies mediocre
    results.  But powerful tools don't imply excellent results any more
    than primitive tools imply mediocre results.  Why should somebody
    be denied the opportunity to create any form of art because they
    weren't born with the requisite motor skills?  If they can conceive
    of the results they want, and can use the appropriate tools to achieve
    them, so what?  What virtue is there in "new, difficult to play"
    instruments other than to randomly single out certain individuals
    who lucked out in the genetic lottery?  Executant skills and creative
    skills have nothing to do with one another.  Cf. Stephen Hawking
    for a stunning example.  We might as well say, "Sorry Steve, your
    physics is junk because you can't write it on the blackboard like
    the rest of us".     
    
    re .21 - I don't think the trend is likely to be toward more expensive
    boxes with longer "shelf lives".   If anything, all the evidence
    points towards the opposite - cheaper boxes with shorter product
    lives.           
    
    I also doubt that the trend will be towards more boxes that "do
    everything".  I expect, again, the opposite.  People don't want
    5 keyboards and 5 (incompatible) sequencers.  They may want one
    master keyboard, one master sequencer and 5 sound generation modules.
    Eventually the new user market will be relatively saturated, and
    vendors will realize that the rest of the money is to be made in
    the "aftermarket", with modular addons.
    
    re .22 - I no ime not usposed too due this, but it's rapport, not
    repore.  ;^)
    
    len.
     

    I can't argue the 'no talent = poor music' argument, and I agree
    that having too much junk can clutter up things (I only want 1
    sequencer, and would prefer that sequencer to be a computer that
    takes care of managing all of the millions of odds & ends that crop
    up).
    
    Powerful tools, however, do put the burden of playing a flashy bass
    pattern on the people who like to generate these sorts of computer
    programs. Just as a point of reference (Let me put on my bullet
    proof vest first), take a look at the Yamaha SHS-10. Available for
    $130-$170, with MIDI output. It has a simple minded sequencer, and
    a chord follow function, with a Bass follow function (either can
    be used of turned off). The drums are not spectacular either, but
    sound ok. The point is, if you want a bass pattern that fits your
    current drum pattern, and the chord that you are playing, you only
    need to press 3 or more notes, to form the chord. There are 20
    different drum patterns, and the same amount of bass patterns.
    While you can say that a real bass player is usually much better,
    you can't say that this concept doesn't work. It will probably be
    expanded upon. For those of you have have not tried the the SHS-10,
    you might want to (try it thru a PA system, it sounds better than
    thru a stereo), and see what I mean.
    
    							Jens   

    re: .31 (Len F.)
    
    > Cf. Stephen Hawking for a stunning example.  We might as well say, 
    > "Sorry Steve, your physics is junk because you can't write it on 
    > the blackboard like     the rest of us".     
    
    I learned a similar piece of philosphy from some something written on a
    stall in the men's room.  It looked something like this:
    
    ---------------------------------------------------------------------
    !									!
    !				PHYSIC's TEST				!
    !									!
    !		Name: Albert Einstein					!
    !									!
    !		Problem 1 answer:  E = MC�				!
    !
    :
    				:
    				:
    
    That part was written in black ink.  There was more writing in red
    ink (the graders marks) that said "Nice job Albert, but next time
    show work:  D-"
    
    This seems very apropos.  There are some purists around who "grade"
    the value of the work based on how much effort or ability was 
    demonstrated, rather than on the content of the work.  For example,
    Keith Emerson's Karn Evil 9, part II (an outrageous piano solo) would
    be somehow less valuable if he had sequenced it rather than played it.
    
    To me, "how" something was done is irrelevant when I'm just listening
    to a piece.  This is not to say that I don't appreciate a
    "performance".  Rather, I just never felt that a "performance" is
    required to call it music.
    
    The creative aspects of a musician are all in his head.  His fingers
    (ok, and his feet for some of us) act as "filters" on his creative
    flow.  In fact, for all the amazing things that music equipment does
    these days, even the best equipment acts as a filter.   My interest
    in other musicians is in hearing their creative ideas and I'm happy
    to remove as many filters as the technology allows.
    
    Sequencers, drum machines, etc. not only allow musicians to express
    their creativity with less filters, it also allows more people to
    pursue musical efforts.
    
    So, for me, the bottom line is let's understand that music isn't
    limited to "performances".
    
    My own radical opinion is that people deprive themselves of much
    musical enjoyment by letting insignificant things shadow their
    evaluation of music like:
    
    	1) It's a drum machine
    
    	2) It was sequenced
    
    	3) It's Michael Jackson
    
    	4) It's pop music
    
    	5) The guy who did it is an jerk
    
    	6) He doesn't write his own tunes
    
    	7) It's a rip-off of _____
    
    etc.
    
    Justice is supposed to be blind, and in my opinion so should music
    appreciation.
    
    	db

    My, we have waxed philosophical, haven't we? 

    My 2� on philosophy is as follows: what's your motivation?  Do you
    listen to music just for a good time, or because you like to jam, or
    because you like to dance, or because you like to analyze it and do
    technical surgery on it, or ...

    I personally don't care for Michael J's music (lyrics we won't
    discuss), but if someone else likes it, who am I to dog him or his
    listeners?  Same thing goes for playing as listening - and yeah, I get
    disgusted when I hear about what a "bad keyboardist Nick Rhodes is".
    But I have the choice (and ability) to change channels or to hit the
    off button. 

    Enough of this for me.

RE: What's Next?

    Funny you should bring up the "system" concept, len.  I've been
    thinking along the same lines.  There may be a few more "breakthrus" in
    the basic equipment like a reasonably priced keyboard controller with
    poly-aftertouch, or some kind of a rack-mount dedicated sequencer &
    workstation thingie, with a user interface at least as good as the ESQ
    sequencer interface.  Other than that, I think that SGU improvements
    and signal processing are going to be the next fad.

    Look in the near term for either multi-FX boxes (a la DSP128) with
    seperate outs for each effect, or cheap smaller FX units (a la
    MicroVerb) with SOFTWARE loadable effects, either via ROM or via MIDI,
    and with around 32 presets.  Kinda like the MT32 or FB01.  Also look
    for more 16 and 32 voice polytimbral modules with seperate outs per
    voice, on-board EQing and FX (a la D50), and better s/n ratios (a la
    TX802).  And of course, more computer-based tools for each of these.

    I think a music workstation is a long way off, but a "workbench"
    concept family of software packages (integrated sequencer, SYSEX
    dumper, patch librarian, MIDI merger, etc) will probably be written one
    of these days.  Dr. T's has the right idea.  I think that the current
    MIDI with all its limitations will be here for a long time.  There may
    be some type of update to the spec to allow more than 16 MIDI channels
    to be defined (or some way of addressing more than 16 instruments at a
    time), and perhaps DELNI-like "merger" boxes (hardcoded MIDI-channel
    mapping) that allow old spec synths to be addressed as post-spec
    devices.  I suspect that the baud rate will at least double (s/w
    switchable), that the current cabling will remain unchanged in the near
    term, and that the DELNI boxes will handle baud rate reduction for old
    synths. 

    As it stands now, a good patch programmer with knowledge/expertise in
    several synthesis styles (ie FM, LA, additive, sampling) and a good set
    of computer tools could stand to make a good deal of money.  I don't
    think there'll be much change in that arena in the next few years. 

    My opinions, for what they're worth.

-b

(Boy, I haven't been able to get to reading COMMUSIC in about a week, so I'm 
sure I look like a "Johnny-come-lately" on this...)

A couple of comments in previous notes in this topic make me put forth a couple
of thoughts I have had over the last couple of years.

One is the use of fiber optics to replace coax to interconnect digitally-
controlled hardware.  As it turns out, using fiber in this kind of application
is *perfect*, even more so than general computing environments.  In general
computing environments, there's a strong (and reasonable) tendency for users to
demand an Ethernet-like approach for fiber-based networking, mostly due (in my
humble opinion) to the desire to minimize the cost and effort for network
reconfigurations and expansions.  As a former network manager of a large
Ethernet-based facility, you can't argue with beauty of just "tapping into"
Ethernet to get another node talking to everyone else. 

But music networks as we've come to enjoy them (;-) tend to be point-to-point
networks, and that's just fine...as long as we get the devices with MIDI-THRU
in the right locations.  Since the single biggest problem with fiber is
fusing the ends to make connections, we get to completely ignore this problem
by virtue of our daisy-chain approach to interconnection.  (I can just see it
now...I walk into Daddy's Junky Music and ask, "Hey Rich, got any 10' single-
mode fiber?  In green?")

Anyway, this rambling isn't all too earth-shattering, except that it ties into
the other comment earlier in this topic: the need for multiple outputs for
such things as "universal DSP boxes."  My feeling is, who needs outputs at
all?

MY dream (ala, "What's Next?") is to have a *totally* digital system.  We're
talking "totally, man", like, no analog signals at all.  Especially no analog
audio signals.  My dream is to have a single fiber connection that provides
both multiple channel signaling (ala MIDI) AND multiple channels for the
resulting audio signal.  All packetized on single-mode fiber.  (Ah, the wet
dreams abound: imagine going *completely* direct to your erasable CD recorder/
player...)

Let's see...a couple quick calculations.  A single, real-time digital audio
track needs 16 bits at 22Kbps (so my dog can hear *everything*) comes out to
352Kbps.  That's a mere fraction of the FDDI fiber standard of 100Mbps (and
most folks think fiber will do MUCH better in the near future).  Actually,
100Mbps is far less than what most fiber will do today, but 100Mbps is a
good conservative figure.  At 352Kb, 24 tracks would end up eating ~8.5Mb.
Oh yeah.  Don't forget about the extra 31.8Kbps for the MIDI signaling...  ;-)

Anyway, I for one, am tired of the bazillion audio signal patches in my
system...and all the constraints that result from it.  For me, digital is the
only way out.

I plan on starting development of such a totally-digital system early next
week...

...just as soon as I win MegaBucks this Saturday nite.

	...jay

    
    I always just wanted to have *everything* wireless.  Guitars, fx,
    keys...everything.  Just have a little antenna to plug into the
    jacks, and then have a bunch-o-dip-switches on the mixer.  And have
    the amps (FET, of course, for weight) built into the speakers, so
    they could be wireless, too.
    
    Just Dreaming
    Ashley

re: .36

Ah yes, wireless is indeed beautiful.

But it makes for *hell* when trying to network.  Recall that when the 
original MIDI "consortium" got together, one of the primary goals was to
make the technology immediately affordable; hence, the puny 31.8Kb bandwidth.

I can't imagine the cost/complexity of the hard/software needed for wireless
networking.  At least fiber is straightforward...

(I'm sure our kids will look back at this note and say, "Sheesh, what a
techno-dweeb.  We've been totally wireless for 5 years now...")

	...jay

    Networking is the ticket... fiber may not be only because of
    the delicacy of the cables. Tha *@$$ things break all of the 
    time in applications where they are not fixed in place and never
    touched. And even then they break all to often. Stick with coax
    it has a chance of surviving your roadies.
    
    Jim
    

    I once designed a fiber link for DEC and ran a 200' cable to a terminal
    for 3 months with no problems.  It was never productized; our office
    was shut down.
    
    Wireless could mean infrared laser diodes pointed at a reflective
    ceiling, not radio.
    Tom

    How about an 'intelligent' sampler ? This beastie would perform some
    type of spectral analysis of a range of notes from a single instrument
    and then use in-betweening techniques (similar to those used in computer
    aided animation) to calculate the correct spectral profile for any note
    *in real time*.

    This would mean that people could add processing speed into the arguments
    over which sampler was best :^).

    As an upgrade you could increase the memory to allow for touch sensitivity!

    Personally I'd settle for a decent (affordable) MIDI guitar controller.


    Ho hum,

    Derek.

    
    
   In terms of all this deliberation I see that Samantha Fox is now
    in the U.S. charts.
    
    Need I say more  ?
    
    						Paul.

    
    I like Samantha Fox,  but not her music.  _sjz.

    re .41, .42
    
    And Wednesday night on KROQ, the guest DJ was Tracy Lords.
    
    She's already an intelligent wireless controller...;^)
    
    Ashley

                                                      
    re -1 
                                                      
    Without wanting to sound chauvinistic I would say that young Sam
    Fox is not hindered by by any form of intelligence controlled or
    otherwise. 
    I am just surprised that you discerning chaps across the water have 
    succumbed to her wit and charm. It can't possibly be the music. 
                     
    Can it ?
    
    						Paul.

    I think in the future, we will all have totally wireless, digital
    Samantha Foxes in our studios, and what else will we need?
    
    (I'm surprised we've succumbed to her charms as well.  I must not
    have been watching when it happened)
    
    Steph

    
    I would say we're going into a rathole here, but somebody would
    accuse me of a chauvinistic double-entendre.
    
    Ash

    The batteries would last quite a while with that much memory 
    capaxity... maybe a year or so...
    
    

    Back to the subject at hand, as it were...
    
    I've been working out a design for an all digital synth.  It shuffles
    numbers around (8 and 16 bit), in multiple representations (positive
    integer, signed integer, positive fraction, signed fraction, fixed
    point, maybe floating point, etc.).  It accomodates subtractive,
    additive, FM, wavetable etc. models.  I spec'ed out the computing
    needs of a 4 voice per timbre, 32 timbre (2 MIDI inputs) (= 128
    voice modules, dynamically assigned of course - the 4 voice per
    timbre is just "average"), 44.1 KHz * 16 bit output rate.
    
    It's a mere 256 MFlops.
    
    Hey, in a couple of years that won't seem so crazy.
    
    len.
    

    Yeah, but Len, is that .25 Gflops in tasks that can be allocated
    
    among parallel processors or in innovative parallel architectures,
    or does it need to be processed serially?  Becauase you shouldn't
    expect mono processors to run that fast.  A josepphon junction
    11/780 would only run about 1000 times faster, but you need
    256000 times faster.  .25Gflops will be reaches with innovative
    architectures. I think one company already claims it.
    but the tasks have to be dividable and configurable for the
    architecture.
    Tom

    It's divided among a jillion separate modules.  I don't have the
    numbers with me; no it's not a monolithic processor, and yes, the
    architecture allows parallelism out the gazoo.  Conceptually, each
    synth component (oscillator, EG, filter, etc.) could be a separate
    module.  More likely, they'd be ganged together for all 128 voices,
    since the voice is the coarsest granule of parallelism.  Some of
    the operations are pretty lightweight computationally (e.g., the
    oscillator requires a few adds and multiplies to scale the phase
    angle appropriately for a wavetable lookup); by far the hairiest
    demand is made by the filters (one multiply and add per section/delay,
    plus some overhead).  Most of the "bigness" comes from the high
    degree of duplication (i.e., 128 voices, with each voice using
    8 oscillators, 8 LFOs (special versions of the basic oscillator),
    32 EGs, 128 maps (to do various forms of scaling for note number,
    velocity, aftertouch, etc.), etc.), which provides numerous
    opportunities for parallelism.  My estimate of .25 Gflops was meant
    to be very conservative.  A more "practical" implementation (e.g.,
    32 voices with 2 oscillators each, with the rest of the components
    scaled down accordingly) could probably be done with today's
    technology, albeit still not as cheaply as current (less flexible)
    architectures.
                            
    len.
    

     EE Times has reported a number of AI projects IBM has been working on, 
     including a music education tutor.  Apparently this is still in an R&D 
     phase.  Here is a quote,  used without permission:


     "The AI-based music tutor, Harmony Intelligent Tutoring System, is an 
     attempt to attack one of the more complex learning tasks using a
     PC-based system.  Harmony, according to Linda Soris, the project
     leader and former music teacher, is one of the key classes in	
     university-level music education.  The tutor integrates an
     interactive,media-rich envoronment (graphics, sound and entry) with a
     formal,logical subject, harmony, which is logically similar to
     mathematics, and can be captured with AI-style rules. 

     "The tutoring  system tracks student progress,tailoring lessons to   
     his or her current knowledge.  Moreover, unlike most tutorial systems,
     it tries to discover the conceptual basis for the student's mistakes
     and to help resolve any conceptual misunderstandings.  Additionally, the
     software is highly interactive, providing instant feedback." 
    

     Notice the obvious acronym for the program: HITS

    This has already been done with counterpoint tutor.  cf.
    computer music journal a little while back
    Tom