Using Your Sequencer As A MIDI System Analyser
Craig Anderton reveals how you can use your sequencer as a MIDI system analyser.
Sometimes you only need to look at your equipment from a different angle to discover an entirely new set of uses for it. Craig Anderton explains...
Quick, now: one of the most useful pieces of test equipment in the MIDI studio is... a sequencer? If you've ever wanted to test a synthesizer's compatibility with MIDI guitar, audition a keyboard's patches quickly, check how rack-mount tone generators respond to notes at the extremes of the MIDI note range, align your tape recorder with the minimum of fuss, make sure all elements of your system are responding to MIDI... (phew!)... or just tune your axe, this is the tip for you.
Sequencers, as we know, are great for recording musical events, but it's a less well-known fact that they can also be turned into powerful event generators, using the world's simplest MIDI programming language: step-time recording. The idea is that you create sequences designed not to play back music, but to generate a stream of MIDI data that automatically puts a piece of MIDI gear through its paces while you sit back and judge the results.
The following sequences (and screen dumps) were created with Passport's Master Tracks Pro and are bundled with version 3.2 of the program, but you can use almost any other MIDI sequencer to create equivalent ones of your own.
The problem: I had just received an Oberheim Matrix 1000 for review and needed to audition all 1,000 programs. Punching buttons for all those presets got to be a little tedious, so I opted for some sequenced assistance. The sequence I came up with (Figure 1) sends out a Program Change command, plays a measure of music, sends out the succeeding Program Change command, plays a measure of music, and so on. It steps through all of a synth's programs rapidly and efficiently.
The Step Editor window shows the notes in Track 1: C, played at four different octaves. You might want to transpose up or down to find patches for bass or high parts. Track 2 contains Program Change commands, starting at 000. Track 3 also contains Program Changes but starts at 001 to accommodate synths that number their programs 001 to 128 instead of 000 to 127. (Note: Do not let tracks 2 and 3 play simultaneously.)
Track 4 is set up to record, so when you find a patch you like, just press a key on your synth, and the note will register on the Song Editor window. If you used Track 3 to send Program Changes, the measures in which you've recorded a note (shown in black on the Song Editor window in Figure 2) corresponds to the number of the program you liked. For example, if measures 5, 12, and 37 are black, you 'flagged' synth programs 5, 12, and 37 as they were playing.
If you used Track 2 to send Program Changes, subtract one from the measure number. In the example given above, programs 4, 11, and 36 were 'flagged', and you have a quick list of the patches you want to use.
This sequence (Figure 2) plays an A=440Hz note every measure, and also tests which instruments are responding to which MIDI channels.
Sequencer tracks 1 to 16 play two looped measures on MIDI channels 1 to 16, respectively, each measure containing a whole note. To tune one instrument to another, set the appropriate tracks to play, and run the sequence. As the sequence plays, tune the other instruments.
To test which instruments respond to individual channels, either solo any one of tracks 1 to 16 and check which synth responds, or create a 'channel sweep' sequence (shown as Track 17). This track generates an A=440 for one measure on channel 1, then on channel 2, then channel 3, and so on up to channel 16. This is useful if you want to do a quick check of a MIDI system and see which channels are triggering which instruments. It's easiest if the measure number is the same as the MIDI channel over which data is being sent.
Instruments set for Omni mode, by definition, will be triggered on all MIDI channels; similarly, multitimbral instruments will also be triggered on more than one channel, depending on their settings.
You may already have a test tone generator sitting in your rack, but if you don't, and need to lay test tones on your tapes, here's another approach. Yamaha's rack-mount synth, the TX802, can generate a sine wave you can use for alignment and testing purposes. The sequence below generates notes appropriate for obtaining 100Hz, 1kHz, and 10kHz test frequencies from the TX802. You will need to adjust the TX802's output level for your particular system.
To set up the TX802:
1. Select a Performance and place the cursor on any voice. Edit the Performance so that the selected voice is assigned to MIDI channel 1.
2. Press Voice Select and, from Utility Menu page five, initialise the Voice edit buffer (this produces a simple sine wave). Do not save the initialised voice.
3. Run the sequence. Because you didn't save the initialised voice, as soon as you select another voice, the initialised voice will return to its original settings.
Regarding the test sequence (Figure 3), programme Track 1 to provide 20 measures of 100Hz tone by playing note G1 with +15 pitch bend; Track 2 for 20 measures of 1 kHz tone by playing note B4 with +14 pitch bend; and Track 3 for 20 measures of 10kHz tone by playing D#8 with +5 pitch bend. Play whichever tone you want to use. When changing tones, stop the sequencer, go back to the beginning, select the new tone, then restart the sequence.
At a tempo of 40 beats per minute, each tone lasts two minutes. At 80bpm, each tone lasts one minute. At 160bpm, each tone lasts 30 seconds. Of course, you can also test response at other frequencies by selecting different notes.
This simple but useful sequence (Figure 4) generates the standard E-A-D-G-B-E notes used fortuning guitars, starting with E1. Each note lasts four measures, which seems about right for tuning, and the entire sequence is looped so that you can verify tuning as the sequence repeats. This is necessary with guitars that use vibrato tailpieces, since bringing strings up to tension may sometimes loosen the tension on other strings.
I found the guitar sequence so useful I also did one for bass that plays E-A-D-G.
Many MIDI guitar controllers are optimised to work in MIDI Mode 4 (Mono mode). For a synth to work well with MIDI guitar, each synth voice should be able to respond to its own MIDI channel and provide a monophonic response on that channel. Furthermore, playing a new note on a string before releasing a previously played note on that same string should not trigger a new attack (legato response).
The sequence in Figure 5 generates one measure of notes on each of the 16 MIDI channels (only 12 are visible).
Measure 1 generates notes over channel 1, measure 2 generates notes over channel 2, etc. As the sequence plays, the notes in the measure should repeat once for each of six contiguous channels.
If you hear the measure once instead of six times, the synth is in Poly mode instead of Mono. If you hear the measure 16 times, the synth is in Omni mode. If you hear the measure more than six times but less than 16, the synth is probably in Mono mode, but some of the voices may be assigned to channels that the guitar synth will not trigger, since most of the time a Mono mode guitar sends data out on only six channels.
The notes in each measure are arranged so that the second note's attack occurs before the first note is released, and the third note's attack occurs before the first and second notes are released. This tests for legato response: if the notes do not retrigger, but the first note simply changes pitch twice, all is well. If each note retriggers, or if the first note continues sounding after the second and/or third note triggers, then the individual strings are responding polyphonically (albeit over a single channel) or retriggering with each new attack. The latter response works well with sounds that have long decays, but is less commonly used with MIDI guitar.
The sequence in Figure 6 steps through each MIDI note from 1 to 127 (it does not play back MIDI note 0), with the measure number equal to the note number (eg. measure 60 plays MIDI note 60). This test is useful for checking how a MIDI synthesizer responds to notes at the high and low ends of its range: does it ignore the notes, double back to a different octave, etc?
Sweeping through 127 notes can take a while; setting the time signature for 2/4 instead of 4/4 shortens the process by half. With a tempo of 250bpm, it takes approximately 50 seconds to sweep through all 127 notes, which seems just about right.
There are many other ways to use test sequences. You can test how synths respond to different numbered controllers, generate a pitch bend change to verify that all your synths bend by the same amount, check for aftertouch response, and more. The tape alignment sequence, for example, could be enhanced easily to use all eight TX802 outputs to feed individual tape tracks.
In any event, start thinking of your sequencer not just as a recorder, but as a piece of test equipment. You'll be surprised how much more useful it will become.
Acknowledgments: Thanks to Ian Gilby, editor of Sound On Sound, for initially turning me on to the concept of MIDI test sequences, and to Jim Johnson, who wrote about live applications of test sequences in his article, 'Sequencing for Live Performance' (March 1988 Electronic Musician).
© 1988 Electronic Musician Reproduced with the kind permission of the publishers.
Feature by Craig Anderton
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!