From gacki@sax.sax.de Tue Nov 29 08:57:36 1994
Date: Tue, 29 Nov 1994 11:45:50 +0100
From: Malte Rogacki <gacki@sax.sax.de>
To: Richard Klasic <rick@latlin.mednet.gu.se>
Cc: analogue@magnus.acs.ohio-state.edu
Subject: Re: Why isn't there a Midi-II?

>> generally a royal pain. The latest version (0.8) was much, much better.
>> Maybe MIDI wasn't completely standardisized (sp) in 1982 but I don't take
>> this as an excuse for the flakiness of the first OS. For example, the
>
>Talking about MIDI
>
>Why isn't there a Midi-II which has much higher data speed. Lets say 
>57600 or something standard. It would made MIDI cards obsolete. And 
>perhaps delays a history. 

"Grandpa, grandpa, tell us about MIDI!"

"Okay, kids, shut up'n listen!"

In the beginning there were certain companies having different protocols.
Let's pick Sequential: They had an own serial protocol running at a speed
of 625 kBaud (! - that's 20 times faster than MIDI!) and used special
locking 4-pin-connectors (as found on the back of rev.3.2 Prophet 5s). Of
course, this was a total overkill since the Prophet wasn't transmitting
much data. Than, after a certain time and with the help of different
companies, MIDI was born. I think credit is mostly due to Sequential and
Roland. Me thinks that Roland was responsible for the lower baud rate and
the DIN connectors. I mean, Sequential had a faster system working, why
should they go back? The answer is pretty clear: manufacturing costs. Why
would someone use those non-locking flakey DIN connectors while the whole
world was using locking connectors (like XLR) in critical places? Again:
manufacturing costs. And because the Japanese companies usually were more
critical with this I think that Roland is to "blame" for this. The
equatation is very simple: To make MIDI succeed it had to be implemented in
as many devices (keyboards) as possible; and hence it had to be cheap.

("Get me a beer, kid!")

Now the next part: MIDI wasn't too slow in the beginning. Most instruments
had a much bigger delay than MIDI can ever build up. I still think that the
bandwidth is sufficient for most applications. However, not for all.
Me thinks it was in late 1986 or early 1987 when the Prophet 2000 was
unveiled. This was the first instrument having implemented the now-common
"MIDI Sample Dump Standard". Although the 2000 had laughable little memory
at first (256 kByte) the designers did realize that MIDI was too slow to
transmit sample data at the usual baud rate. So they implemented a
"MIDI-fast"-mode that did operate two times as fast. However, the reactions
from the various MIDI commitees were not very friendly (hearsay), me thinks
mostly because Sequential may have had developed this secretly. So this was
turned down immediately and the Prophet 2000 remained the only instruments
with two MIDI speeds (normal and fast).

But this didn't solve the problem. There have been different attempts at
different stages to get rid of the problems:

(Warning: little analog content from now on!)

Emu uses for both the original Emax and the EmaxII the RS422 bus to
transmit and retrieve samples to and from a computer (special cable
required). Significant speed improvement (hearsay).

Peveay developed the SMIDI protocol. I only know that it can be used to
transmit sample data. I don't know if it can be used to transmit note data
and controller data, too. SMIDI is implemented in very few instruments so
far. If it works, it works REAL fast.

Akai used an own SCSI protocol (not SMIDI) to enable sample transfer
between computers and their S-line samplers (S-1000 and upwards). It worked
very well and was REAL fast. Hint to all Akai users on this list: It is
nessecary to have BOTH MIDI and SCSI connected to the computer since all
mapping informations are transmitted via MIDI (only raw sample data on the
SCSI line).
Maybe Akai pushed this concept beyond that with the EX-1100 (the expander
to the S-1100). My sources report that even note informations were
transmitted from the S-1100 to the EX-1100 via SCSI while controllers had
to take the MIDI cable.

Lone Wolf developed the legendary MIDItap systems that can be used to
transmit MIDI data (and much more; the bandwidth is significantly higher
than MIDI) over longer distances via glass faser links. I've never seen
this in action but Keith Emerson is said to use this system.

And of course there are they rumours about the ZIPI system (developed
primarily from Zeta, a maker of exotic controllers like MIDI violins).

"Gee, kids; grandpa is exhausted."



Malte Rogacki gacki@sax.sax.de  100116.154@compuserve.com
-----------------------------------------------------------------------------
"Don't forget to TURN ON THE SYNTHESIZER. Often this is the reason why you
 get no sound out of it." (ARP 2600 Owner's Manual)
-----------------------------------------------------------------------------

