Conference napalm::commusic_v1

    You ought to seriously consider getting some form of MIDI switch.
    Roland makes two nice little boxes, the MPU-104 and MPU-105,
    that provide switched 1 in 5 out and 5 in 1 out multiplexing and
    demultiplexing.  The 1 in 5 out is basic a switched thru box;
    you can drive up to 5 MIDI ins in parallel with it.  The 5 in
    1 out is a one-at-a-time switch.  The two units can be rack mounted
    side by side.  The thru box is about $60, the input switch about
    $80, and the rack kit about $20.
    
    The JX-3P has no MIDI out?  Extraordinary!

    len.
    

    Not to revert back to note 261 but...
    
             The JX3-P does have MIDI out. It's just not part of my
    configuration. It DOESN'T have MIDI thru.
    
    Edd

    I recently got a 5 in 5 out MIDI switch from a mail-order company
    in Oregon.  I have some problems with the lack of a rugged box
    and the EMI, but it does work.  Saves a lot of cable pushing, and
    makes the distance from the sequencer to each instrument equal.
    When more than one input is enabled for a particular output it
    performs the "merge" function.
        John Sauter

    
    	Sorry to butt in, but this doesn't merit a note of its own,
    and does have some relevance to MIDI setups.
    
    	Does the length of a MIDI cable have any relation to the
    problem of delay?
    
    	I.e., is a 5 foot cord preferable to a 10 ft cord?  Ideally,
    should all the cords be the same length?
    
    						- Karl

    Skew across multiple cable lengths is not a problem, but "smearing"
    of the edges is a problem with long cables.  I have heard over and
    over again, "don't use  MIDI cables over 10m (30 ft) long".  I think
    the standard itself says something about maximum cable length.
    For this same reason, shorter is better; the problem is not the
    skew that Tom mentions but waveform distortion.  For similar reasons,
    only use real MIDI cables, not DIN cables (even though the connectors
    are the same).  The MIDI cables use twisted pair internally, a DIN
    cable is shielded audio cable.  DIN cables also have more conductors
    than is necessary, which the Roland DIN sync convention uses (in
    addition to the "data" line, there's a start/stop line).  At least
    one manufacturer (ProCo?) makes a cable called MIDI-Plus which
    satisfies the MIDI loading/distortion requirements but also has
    the additional wires for Roland sync use.

    The MIDI bit rate is 31.25 Kbits/sec; this is 32 microseconds per
    bit, so Tom's skew estimate is very conservative.
    
    Re John's 5 in 5 out switcher - merging is a function that most
    switch builders avoid by making the input a 1 of n choice; merging
    requires maintaining the time coherence of MIDI messages (i.e.,
    you can only interleave whole messages, hence must provide some
    buffering and data stream interpretation).  A 1 of n input switch
    can be dumb.  I have yet to find a need to merge two MIDI data streams
    in real time that cannot be addressed by the overdub capability of
    my sequencers.
    
    Incidentally, I'm using an 8 in 10 out switch from JLCooper, with
    1 in and 1 out each feeding a Roland MPU-104 and MPU-105.
    
    len.
    

    It's worse than interleaving messages.  You must also watch out
    for the running status byte.
        John Sauter

    
    
    FYI, the MIDI spec mentions 50 feet (15 meters) as the maximum cable
    length. Maybe you could use a line driver/repeater if you need longer
    runs.
    
    I agree with Len, it's not the skew, since you have ~30 microseconds
    to play with. The spec mentions 2 microseconds rise/fall times on
    the signal transitions.......the additional capacitance of a longer
    cable will surely effect this and cause MIDI waveform distortion.
    How tolerant the MIDI reciever is of this I'm not sure, they most
    likely use some form of Schmitt trigger which can tolerate some
    hysteresis.
    
    my $.02 worth
    
    CdH
    
    

    Distortion can also become a problem if you go through too many
    successive THRUs.  Recommended practice is no more than 2!  The problem
    here is that the spec I believe requires opto-isolation for the
    THRU (sort of like a repeater), and the delays and distortions can
    add up.
    
    MIDI protocol and line delay is a problem that is more theoretical
    than practical.  Many synths take longer to respond to a note on
    message (or a keyboard event) than the length of a whole chord's
    worth of MIDI messages.  I have heard numbers like 10 milliseconds.
    At 30 usec per bit, and 11 bits per byte (including the start and
    stop bits, as it's an synchronous protocol), that's 330 usec per byte,
    and at 4 bytes per message, that's 1.32 msec per message, so you can
    transmit almost 8 messages in the time it takes some keyboards to
    respond to one.
    
    len.
    

    I might be interested in buying a switch like that.  Could you let
    us know the company that you ordered it from and the cost?  Also,
    what's the maximum number of inputs that can be merged to any one
    output?
    
    Do you know if this was ever advertised or reviewed in Keyboard?
    This switch sound more versatile than any I've seen there.
    
       Joe Meyers

    See note 304.  I was mistaken about the number of inputs and outputs.
    It's 4 and 4.
        John Sauter