Powerbook 1xx MIDI summary

[This article was written to analyze the MIDI performance of Apple's early Powerbooks. It details a number of problems that occurred under OS 7 - often the only OS these machines could run. I've kept this document mostly for historical reasons; obviously most people would be running a much newer machine with OS X nowadays. - JM 01/08]

This letter contains the sum of Atomic City’s current knowledge of MIDI with respect to the Apple Powerbook computers, containing the findings of recent tests of the 100, 140, and 170, combined with late-breaking new results. We hope you find it useful.

A. THE POWERBOOK 100

The 100 works fine if you keep in mind that it’s only a 16MHz 68000; Finale can make it choke. Old versions of Alchemy won’t boot, but that’s not a PB problem. The “big name” software we were able to try (Performer, Vision, Cubase, UpBeat, Galaxy, and MAX) runs fine. It’s a good idea to disable Appletalk, and have software up and running under MIDI Manager. (Actually, on the 100, you MUST disable AppleTalk, as it will grab and munge the only available port for MIDI, the printer port.) If you use Opcode stuff, OMS goes a long way toward deglitching your work. We weren’t able to test everything in the universe (the test occurred as part of a recording session, not as an end in itself, but that fact got trimmed from the article in EM), but here are some highlights….

  1. The older version of the MTP desk accessory won’t run, but the MTPII utility, which runs under MIDI Manager, should be fine. In Nick’s words: “The MTP-II DA will run with MIDI Manager, so using that (via OMS for example) should be OK (and this is the main reason that I’m considering buying an MTP-II). I don’t know if the MTP-II DA runs on a PB100 using its own drivers. Oh, unless I’m mistaken, the MTP-II software is a proper application rather than a DA; but I could be wrong.” The PB100 can run reliably at MIDI speeds – just. Any faster communication speed (such as MTP’s fast handshake) will not work reliably. To be honest, I’ve never (in two years of MTP use) notice any improvement in using fast handshake.If that limitation is understood and nothing faster than ordinary 1MHz mode is attempted, the PB100 printer port can handle the MTP, MTPII, and any Opcode interface up to but possibly not including the Studio 5.
  2. MAX handled a patcher designed on the IIci, one that was designed to kill portables (and really hung up an Outbound 2030 badly), on the PB100 without breaking a sweat. Nick points out that MAX performance can be sluggish on the PB100 with really big patchers, but that’s true of any Mac– you just have to determine the limits of “really big” by trial and error with your particular platform. I should point out that Nick uses MAX as a controller-data remapper for live performance with CASSIEL, and places tremendous burdens both upon the code and the computer running it.
  3. UpBeat ran like a dream.
  4. The worst timelag we got with Performer pumping eight MIDI streams of arpeggios through the MTP from another Mac into the PB was about 2 milliseconds. Not a major issue.
  5. Galaxy handled input and output of bundle data from the worst synths in our rig, the Roland D-70 and Korg Wavestation EX, without a hiccup. Repeated attempts could NOT produce a failure with ANY synth SysEx dump. We tested Oberheim, Sequential, Korg and Roland gear without a hitch.

Vision and Cubase were not tested as extensively as Performer, but they run, do not lose input data, and no noticeable timing glitches were found, although the torture tests weren’t as heavy-duty as the one for Performer. I am currently using Vision on my PB100 with an Opcode MIDI Translator interface, and it works like a charm.

B. THE POWERBOOK 140, 145, AND 170

Nick Rothwell owns a PB140, which he only considered reliable for output (it lost a LOT of data on input, not just long sysex dumps. Apple was wrong on that point, as on many other points) until a recent Tech Note suggestion by Apple was augmented and implemented by Opcode in a new version of OMS, 1.2beta. Under 1.2b, the input problem is solved, but incoming data loses its timestamping. We’re still testing just how bad a problem these 6 msec hiccups can be. NOTE that this fix only works for interfaces attached to the Modem port! The printer port is and always will be unreliable, apparently, for input. In his own words, Nick says: “I haven’t hammered my machine heavily yet (and won’t be doing until the next live project I have to prepare, in which case I’ll be using both PowerBooks), but the tests I did (including SMPTE lockup) suggested the PB140 is OK. 6msec isn’t too bad, but you can hear/feel that length of time.” The 170 should have no differences in performance than the 140 or 145.

C. FUTURE DESIGNS

According to the latest technical data, the 160 and 180 will suffer from the same problems as the 140, 145, and 170. There is not enough data available to render a judgement on the Duos. We are not authorized beta testers of these machines, and cannot promise any early results at this time. If we do learn something of use, we will see to it that the Editor of EM, who has been very kind and helpful in his efforts to relay our findings to a larger audience, will hear of it and pass it along. We hope this helps possible Powerbook buyers with any decision they make. Our findings neither carry nor imply any guarantees from ourselves or any of the manufacturers involved (especially Apple!). We would, in closing, like to compliment Opcode on the excellent reliability of OMS on the Powerbooks, and to thank Doug Wyatt for the 1.2beta patch that renders the costlier Powerbooks at least partially reliable.( We would also LIKE to express our opinion of Apple’s handling of the whole MIDI question, Powerbook and MIDI Manager support included, but we’ll refrain for legal reasons.) Please feel free to write or email me with any comments; we do not plan any addenda to this study at this time, but circumstances may change.Dr. Michael MetlayAtomic City Productions[out-of-date contact info removed - JM 01/08]Many thanks to Dr. Nick Rothwell of CASSIEL for his technical expertise, and to Mr. Eirikur Hallgrimsson of DEC for his input and encouragement.