Conference napalm::commusic_v1

    I've never used one, but I understand that there are MIDI-controlled
    switches available--at least I've seen them advertised in magazines.
    According to the ads, they listen for a program change on their
    channel.  In response to a program change message they will send
    out what ever program changes you need, on any channels.  Sort of
    like a definable key in TPU.
    
    I don't know how you program these things; I suppose they have the
    traditional hard-to-use front panel, and incomprehensible
    documentation.  But once you have the programming done you should
    be able to hit one button on your keyboard and generate any list
    of program changes within the box's memory limits.
    
    The box has a memory for each of the 128 program change messages,
    of course.
        John Sauter

    
    KM    
    
    2 Questions.
    
    Are you saying that it was impossible to hit patch 44 or just that
    it would have taken to many keystrokes in the availble time ? 
    
    Or are you saying that you had both synths slaved to the same midi
    channel and required different patch numbers for these?
    
    And....
    
    How the hell did you get the KX88 out of the house into the transport
    and at the other end still have the  energy to play the thing?
    Last time I gigged with the KX88 I swore never to do it again unless
    the bucks were to big to refuse and would enable me to hire a lifter.
    But then did you ever try getting a KX88 up and down a loft (attic) ladder.
    
                                Paul
    

    
    Hmm. I want one too. Welcome to 'live' reality, KM. 
    I want:
    
    1. More than one program change OR roll-your-own system exclusive
       (lets think about the future! You have all sortsa stuff out there
        needing different packets, even the FB01s need more than
        just program change) going out on any of the channels
    
    2. one-press activation. Maybe even more often than once a song...
    
    Thats easy. The only product that seems to be able to do this is the
    AXXESS MAPPER. Changes any midi event into other(s). BUt its not
    cheep! If there were a way to get the 'change packet bytes' into a
    sequencer, that would work. Sounds like computer time, and who 
    knows what software lets you format midi messages. 
    
    If I designed,built, and supported a box like this, how-mucha-you-pay?

    ron
    

    
    
    Sigh.
    
    I had the same frustrations the first time I experienced using more
    than 2 midi devices. If only MIDI had been properly designed in
    the first place ...
    
    I sat down and drew up the software spec for a switch box which
    would be flexible enough to keep me happy (I've spent some time
    working on networking software in the past ...). Then I got together
    with one of our hardware engineers to keep me honest.
    
    When we finished we had the design for a slick box which the AXXESS
    MAPPER doesn't even touch for features, ease of use, etc. Our target
    was from 16 to 64 midi node connections.
    
    Have we built it?
    
    No.
    
    In order to do it, and still have time and strenght to play music,
    you really have to go the PC route - particularly for the DIN
    connectors. You also want to go this route for reliability -
    particularly for performance use.
    
    We have all the tools for the layour and circuit verification (after
    all, it would be a lot simpler than what we are building out here in
    the wilds of WA), but neither of us are into building products in our
    spare time. As a matter of fact, we haven't had much spare time since
    we did the design work last year. 
    
    Doing it as a product was the only way we could justify the costs
    of a full fledged board. (Yes, I know, we also can't use DEC facilities
    to make products.)
    
    Maybe we should look at it again if you people are really all that
    interested.
    
    At least doing the design got rid of some of my frustrations.
    
    Robert
    
    

    How could it get rid of frustrations if you didnt
    build it?
    
    You sound very frustrated. My personal view is that
    you overkilled whatever you were trying to design.
    You never really said what problem "it" solved...
    
    Look. Were talking 90% software, 10% hardware. Theres
    really no magik to midi hardware. So why design another
    micro to run this 90% software?  Seems to me the only
    restriction is that the micro be able to keep up with midi
    input stream(s), modify, and ship it out....in the general
    processing case. Pick a system  and write some code....
    Unless you are 'really' going into business and want to
    produce a dedicated application-based micro...
    
    Tell me what you want and what it's worth to you....
    
    You want a PC board for a DLV to midi interface? I'll be
    anouncing that next week. Use a cheep 11 or a not so cheep
    uVax for that processor...
    
    ron
    

    
    
    Well, I agree the hardware in terms of the micro itself is fairly
    simple. The duarts and optocouplers are fairly simple. The wiring
    is a real pain in the @##.
    
    The problem we looked at was to define a flexible network between
    up to 64 midi devices. It was intended to allow the creation of
    sub-networks of devices. Of course, it featured full midi merge
    for multiple sources.
    
    A "device" here means a midi channel from a real device.
    
    The design seemed to satisfy all my requirements. Of course, the
    patching was all midi controlled itself with the switch box just
    being another midi device accessable through a reserved (but
    extendable) sub net.
    
    I suppose, if you really wanted more details, I could find the notes
    somewhere around.
    
    The main problem with building it was just the shear wiring of all
    the input/output connectors, as I said.
    
    I might get around to doing it anyway. My time recently has been
    largely consumed by building a 30 x 16 foot studio. It is almost finished
    and is completely wired for almost anything (video, audio and midi
    - outlets every 4 feet). So now I have the wiring/switching nightmare
    that we did the design to solve ...

    My main frustration (as I've said elsewhere in this notes file)
    is with the braindamaged midi design and people like Jim Cooper
    who seem to be responsible for it. However, I'd rather have a
    braindamaged midi than no midi at all ...
    
    Robert

    

    It's clear that midi is not a very good NETWORK protocol.
    I dont want to debate if it *should* have been or if it
    even *could* have been given the designers backgrounds.
    
    For all the people griping out there, there are many of us
    that are making midi work in musical and computer contexts
    with few problems and more than 2 'nodes' in the connection.
    
    I find it strange that after having problems and frustrations
    with 2 midi devices and the 'shortcommings' of midi itself,
    that you set out to design a 64 device "flexible network".
    
    I mean, if midi was so bad with 2?      64?
        
    Are you performace or composition oriented? What'dya planning
    for the big studio?
    
    ron
    
    
    
    

    I'm currently using five synths, three drum machines, two MIDIed
    reverbs, and two sequencers with no problems.  All at the same time.
    I route everything through a JLCooper (yes, the source of all evil)
    MSB-1 (8 in 10 out) switch box and a Roland MPU-104 (1 in 5 out)
    and MPU-105 (5 in 1 out).  I can only play one keyboard at a time
    so a MIDI merger is of no use to me.
     
    MIDI may be "brain damaged" (I seem to be hearing this word a lot
    lately to describe successful products), but it works for me.
    
    Could you perhaps share your observations on how MIDI is deficient?
    
    len.
    

    Back to my original performance-oriented gripe, yes, I was driving
    2 synths from a KX88, which can xmit only on two MIDI channels at
    once. I thought I was slick in figuring out how to set up a foot
    controller for MIDI channel 1 ONLY, so I could bring the Kawai K3
    up and down in volume relative to the Roland MKS-20 (chan. 2)
    
    The problem with the KX88 is that the 16-voice 'bank' voice selection
    maps to decimal very poorly ! And now that I have an FB-01 with
    48-voice banks, the problem is even worse !
    
    I see all kinds of MIDI foot-switch devices advertised, mostly quite
    reasonable in price.. as I DO have a MIDI- MACintosh, of course
    the 'issue patch change' problem would be gone, BUT I don't use
    sequencer software for every piece, and I certainly don't wish to
    GIG with the MAC ! (That's what the sequencer in the EMAX is for.)
    
    Len, could you post a short reply stating what each of the MIDI
    switch devices you have actually DOES for you ? 
    
    best, karl

    Essentially, the MSB-1 is a crossbar switch and I use the MPUs to
    multiplex one each of its ins and outs. The MSB-1 is 8 in 10 out,
    the MPU-104 is 5 in 1 out, and the MPU-105 is 1 in 5 out.
    
    The MSB-1 is my main MIDI "switchboard".  It takes 7 ins from the
    MSQ-700 sequencer, MC-500 sequencer, SBX-10 sync box, JX-10, CZ-101,
    Juno-106 and the MPU-104 (1 input remains available for future use).
    The 10 outs go to the MSQ-700, MC-500, JX-10, CZ-101, Juno-106, MKS-80
    (Super Jupiter), MIDIBass, SBX-10, and MPU-105.  The MPU-104 selects
    one of the Octapad, TR707, TR727 and TR909 as a MIDI source.  The MPU-105
    is a "switchable" thru box that can drive any or all of the TR707,
    TR727, TR909 and two SRV2000s.  With the turn of a knob or push
    of a button I can route just about anything anywhere.  I.e., the
    MSB-1 has 10 knobs, each of which select for that output 1 of the
    8 inputs.  The MPU-104 and MPU-105 each have 5 buttons.  With the
    overdub facilities in the sequencers, and given the way I generally
    sequence things (i.e., one part at a time, by myself), I have yet
    to need any merge capability, as I never need more than one MIDI
    source at a time routed to the *same* input.  Note that I *can* drive
    all the synths from the sequencer(s), while driving the drum machine(s)
    from the Octapad, if I want to "play along".  I can also overdub from
    the Octapad into the sequencer while it is driving the synths. 
    This is basically a studio, as opposed to live performance setup.
    It provides no functionality beyond basic connectivity.  Everything
    else is done via programming.
    
    In the normal configuration, everybody listens to the MC500, and
    the MC500 listens to the JX-10 as master keyboard.  The MC500 "soft
    thrus" everything from the JX to everybody else.  The JX runs
    "nonlocal", i.e., its synth modules are not connected internally
    to its keyboard.  Would that Roland had made it possible for the
    JX to send on different channels than it receives on (it sends/receives
    on two arbitrary channels, and the patch storage can send program
    changes on two other channels as well); then it would be a fully
    general master.
    
    len.
    
    
    

    Thanks, len. I extracted your reply, and am gonna think about its
    implications for my setup.
    
    karl

    



Re: .8

>                -< Yeah, But *I* Would Have Done It Better... >-

Len, I'm sure you woul have done it better, I'm not sure what "it" is, 
but I have confidence that you would have done it better. :-)

>    MIDI may be "brain damaged" (I seem to be hearing this word a lot
>    lately to describe successful products), but it works for me.

Right. Being "brain damaged" doesn't make a product unsuccessful, just a
pain in the @## to use - at time. I'll bet few people would disagree with
that statement with respect to MIDI.

MIDI works for me. It also causes me pain and bites every now and again.
A better original design might have helped.
    
>    Could you perhaps share your observations on how MIDI is deficient?
    
I've made a few comments elsewhere in this notes file. (Everytime I've
complained, you've generally jumped on them ... :-) ).

Sure, I can give a few comments about why I think MIDI is deficient. Before
I do, I'd like to point out to you that I am perfectly aware that the more
complex something is made or the more functionality that is provided, the
higher the cost. As I said to you somewhere else in this file, having built
a number of successful software products (and some hardware ones), I'm well
aware of the cost of development and the need for return on investment.

(drags soapbox out, tries to remember to be brief ....)
    
Deficiencies:

* bad topology: without any pass through requirements for each device, this
  is what causes the need for merge and through boxes (and makes JL Cooper
  his income). the connectivity is my main beef with MIDI.
  an ethernet type of approach using RS 422 would have been better in the
  long run. the protocol would have been different (due to possible addressing
  requirements) and costs higher. however, it would have the benefit of
  possible software configuration - currently one has to "move the cables"
  or buy enough routing boxes to make it work (KM's problem at performance
  time).

* a bit oriented protocol: this is primarily due to the line speed (ie.
  pack the information as tight as possible). this limits future possibilities
  for expansion.

* 31.25 kbaud when the rest of the computer world is different. this is
  a "little" sssssslow.

* non-acknowledged messages (primarily for speed) - causes note off problems.

* 5 ma current loop. nicely allows for protection against miss-connection,etc.
  but little else.

In short, it can be a pain to configure a MIDI network from time to time.
That problem could have been avoided in the original standard. Even if the
current method of connection was used (for cost cutting), the protocol could
have been designed for something like RS 422 which could have been available
as an option (ie. higher cost), or when costs fell.

As it is, I suspect that after MIDI 2.0 (largely concerned with inter-sampler
message formats) there will one day be a MIDI X.Y which will work over a
different hardware interconnect. Then we'll have V1.0 to VX.Y adapter boxes
(more products for JL Cooper).

Just like a computer company you say? (Unibus, Massbus, BI, NI, CI, Q-bus
and the future ones I can't mention in a notes file ...)

Len, in your reply .10, you gave a very nice studio layout of a MIDI network.
You pointed out that you've never needed to merge midi streams. I have and
I know others who have too (like midi recording multiple devices at the
same time). There are ways of avoiding these circumstances, but they all
involved extra things which particular devices do automatically (like automatic
pass through). That is, the solutions involve a particular manufacturers
facilities, not facilities provided in the standard.

In .10, you also point out one of these facilities you would like in your
JX. Such things could (added expense acknowledged) be provided for by the
standard.

These days when I look at a new piece of electronic MIDI gear, one of the
first things I do is go over the MIDI implementation and see what of its
functions are accessible via MIDI. I also look for automatic pass through
capability. Such features colour my opinion of the instrument a lot.

I am glad that MIDI exists in the form that it does. Finally electronic
instruments can be interconnected and we all have seen the benefit of that.
I do get mad (really mad) reading the crap that Jim Cooper continually writes
in Keyboard - he's gone over the same basic points for a longggggg time
now. He also avoids issues left, right and center. I find it hard to divorce
the MIDI spec from him. (Next on my list is Steve de Furia and his idea
of software engineering in PASCAL ... but that's another story - at
least he is trying to present things for musicians.)

So far I've targetted this reply at you, Len, because you always respond
quickly and decisively. On the beat, so to speak.

I suspect you'll respond to this and tell me I that I'm every bit as dumb
as I already know I am. 

I don't have a big axe out for MIDI. It's a tool for me and I like my tools to
work well and work professionally. However, it really doesn't matter how good
the tool is if the workperson isn't skilled or talented. I prefer to spend my
time in practice and composition, not dealing with the problems of MIDI 
interconnection. 



On to a different topic:

Re: .7

Why 64? Why not?

I guess the answer is wrapped up in the size of the studio ...

I'm lazy. I like to do things once and when I've decided to do something
I work out what I'll need in the next 5 years (as best as I can guess) and
work out how to achieve that. The ability to interconnect 64 MIDI sources
would certainly hold me for that length of time (probably longer). So will
a 16 x 30 foot studio.

The studio won out over the switch box for my limited "free" time:
    
Building a largish (for one person) private studio was not that much more
trouble that building a small one. I got tired of having MIDI and mike cables
running all over one floor in my house. I got tired of tripping over guitars,
dropping wind instruments on the floor (and then stepping on them) and dealing
with my two cats making their beds in from of my monitors (they are probably
both related to Len's cat ...).

I also got tired of waking people up at 2 am. (Actually, more tired
of dealing with woken up people at 2 am ...)


So now I have a studio that I could put a small band into and not bother
anyone (either living/staying in my house or in other houses on the street).

That's what it's for: increased ease of creating music.

Now, one day I might get up the courage to make a submission to a commusic
tape ...

Robert

    like you say, "your axe isnt that big" !.
    
    as far as I can tell....
    
    ron
    

    Robert, I was responding more to the tone of your comments than
    their content, which was anything but dumb.
    
    All your points are valid.  But like all complaining about standards,
    I feel those of us who "know better" bear some of the responsibility for
    what we've got, if by default.
    
    The guys who designed MIDI are not network jocks.  They're not computer
    jocks either.  At best, they may be competent electronics types.
    (I mean, Cooper has built boxes that work; I'm not cynical enough
    to believe he *deliberately* screwed up MIDI so as to have the
    opportunity to build lots of such boxes and get rich on it.)
    The only way we're gonna get decent standards is to help write them
    ourselves.  Do I have the time or the inclination (I *am* involved
    with international standards work for Digital) to do this?  No.
    Given that, do I feel comfortable pissing and moaning about MIDI's
    deficiencies?  No.  Should I write to IMA and tell them what they
    should be looking out for "next time around"?  Yes.  Will I?  Probably
    not.  Note that I *have* transmitted to Roland (via my buddies at
    Wurlitzer, who forwarded my comments to the Roland sales rep with
    the admonition "listen to this guy, he knows what he's talking about";
    better than a blind letter to corporate headquarters) my satisfactions,
    wishes and disappointments re their SRV2000, MC500, Octapad and JX-10.
                                                      
    Yeah, Cooper's and DeFuria's columns in Keyboard are content free
    to anyone who knows what they're doing.  But do you have any idea
    of the average level of competence with respect to computer programming
    and network protocol architectures in the musical community catered
    to by Keyboard?  Would you believe ZERO?
     
    This conference has some pretty sharp people who are really at the
    cutting edge of musical applications of digital technology.  We
    take the level of sophistication of *this* community for granted
    (have you seen some of the garbage that parades as informed opinion
    in net.music.synth, the USENET equivalent of COMMUSIC?), but are
    probably "representative" of the top 0.1% of folks working in this
    space.  So why don't we propose a next generation MIDI?
                   
    The title of my reply (.8) was my tongue in cheek way of making
    this point more concisely.  I get a little weary of people who
    won't put in the work complaining about what other people have done
    for them (no cheap shots aimed at present company intended).  If
    we really believe we could have done it better, than by gawd, let's
    do it better.  This conference is a perfect forum for drafting such
    a proposal.  I once tried to get such a discussion going on the subject
    of pc-based sequencing software, but there weren't any takers.
    
    It *is* much easier to complain about what somebody else did than
    do it yourself, I'm guilty of this myself (I'm not building my own
    synth in my home workshop, nor am I writing my own sequencer software
    for my Amiga - I just complain about the lack of decent musical
    software for my favorite pc, but I have a notebook full of
    requirements).
    
    Complaining about MIDI amongst ourselves isn't going to help anybody,
    except possibly to make us feel that much more erudite.  And
    complaining to the IMA isn't going to help much either - they can
    quite legitimately say "well, what would *you* do?"
    
    Can we do it?   Can we come up with a draft next generation proposal
    that's economically practical and exploits the technology?  And
    it's *got* to be backwards compatible with MIDI - there's just too
    much gear out there to just obsolete the standard.  A new standard
    that doesn't address backwards compatibility simply can't succeed.
    An unpleasant but no less true fact.
    
    Anybody feel like bellying up to bar?
    
    len.
    

    >  ...bellying up to the bar?
    
    Ah, *hic*, yeah, ah cow me in...
    
    Ed*hic*d

    [set/inflection=cockney_1]

    Aye. Volunteers? Aye. Do. Eh?
    
    [set/inflection=BASton_eze]

    Why not. We need something to discuss of import at the
    wednesday meetings. Ah, and some way of recording the
    brainstorm.
    
    Count me in
    ron.

    Remember the topic, MIDI switchers ?
    
    I've been doing some store-hopping. Both Yamaha and Roland offer
    floor mount MIDI stomp-switching. Unfortunately both of the units
    only offer a big *8* sets of changes. The Yamaha is apparently meant
    for guitarists who wanna switch their SPX90 remotely. How many
    guitarists own SPX90's ?
    
    karl

    
    In terms of the original problem which I think was relating effects
    to synth patches. I think I may have solved half the problem. I
    have on loan at the moment from he-who-my-wife-hates-the-most. A
    midiverb II. It's pretty good. It has more effects than the midiverb
    including chorusing ddl flange etc. And these can be assigned to
    any midi patch number. The basic sound seems to be much better (to
    me) then the MIDIVERB I. And at about twice the cost of a
    midi-controller could possibly be an option.
     
    
    

    wha???
    
    what does THAT have to do with a midi-switcher?
    
    rr
    

    re -1.. check out .18 where PK says the MIDIverb II's effects can
    'be assigned to ANY MIDI PATCH NUMBER'. This means that one could
    match effects to patch and change them (on the same MIDI channel)
    simultaneously. This precludes (at least for effects switching)
    the need for a programmable pedal switcher unit.
    
    Except I have a MIDIverb and a new MIDIfex ($199). naturally.
    
    8AC9/karl