but of course my question here + now is:
was MacOS8/9 able to have more than 32Midi Ports? (4 x 8 = 32)
the MOTU MIDI Timepiece manual also makes refernce to networking multiple MTPs of course; was that also up to a max of 4? if so why this max of 4? is there a real limit to the amount of MIDI ports a single mac can address? (32?)
The first thing you have to do to solve this is define terms.
i.e. What exactly is a "MIDI port"?
If we define a "MIDI port" as the actual 5-pin DIN connection (passing for the moment on USB and that whole world) we must remember that each instrument / device connected in this way may require
two of them - 1 in, 1 out.
The Mac (and PCs too, for that matter) do
not have "MIDI ports". They do have serial or parallel
data ports that can be used to handle MIDI data, provided there is a suitable
MIDI interface connected there.
These ports cannot be parallel combined or daisy-chained like SCSI so the maximum number of devices they can "see" is one per port. The fact that you can connect more than one UM880 or Opcode Studio is due to the interface's ability to combine and route all their data that presents to the computer as "one" device.
So, the number of
interfaces that can be connected at once depends on the interfaces themselves and the maximum data handling capability of the computer and its port(s).
By that I mean: While MIDI interfaces such as Roland, Opcode or MOTU units can be stacked / combined in such as way as to create a LOT of MIDI ports, there is then a practical limit as to just how much data you can stuff through the serial computer port(s) at any one time. There's also an issue with how many unique ports you can
address at once which is where OMS and FreeMIDI come in but that's a different worm also.
The total amount of MIDI data flowing to and from the computer at any given moment is of course, one bit…lol.
The total amount of MIDI data flowing to and from the computer at any given
time is highly variable and depends both on:
1) The number of devices connected.
2) How many are active.
……and finally
3)
How complex the data is.
Obviously, a simple say, drum pattern sends a LOT less data than a two-handed piano part or a complex synth part with continuous controllers and pitch bend. In a given sequence, the
data rate varies constantly depending on the track count, music genre, types of instruments used and a dozen other variables.
Just because you can connect
x number of ports at once doesn't necessarily mean you can really use them simultaneously.
There is a maximum amount of data you can pass and process in real time but there is no fixed number of ports that can "equal" that rate.
Opcode, for example, suggested you can stack six Studio Fives - two groups of three on each of two Mac serial ports for a total of ninety unique in/out MIDI pairs.*
Opcode also made a point that the MIDI data stream can easily get "clogged" at a certain (unspecified) point and may require "thinning" measures to be taken and/or the port data rate adjusted to keep things flowing because the serial port has a maximum transfer rate
and the computer has a processing limit dependent on proc clock speed, software quality and a dozen other factors.
So…There are no "official documents" or specs regarding this because of the infinite variations of individual setups. It like: how much traffic can you handle on a freeway? It depends on the speed limit, the types of vehicles (cars vs trucks) number of lanes etc etc. You can only estimate a
nominal amount. Now, translate that to MIDI hardware / interfaces / ports. Opcode, for example, seemed to pick 6 x 15 = 90 total as a practical limit.
Conclusion: This is clearly a classic example of "Your mileage may vary"*Each one of those can be theoretically multiplied by 16 channels for a total of 1440 addresses.
I seriously doubt anyone has a real use for that many destinations of that an OMS setup window or a SVP instrument window could be readable with that much content.