Make your own free website on

MIDI patch librarian (click the name for downloading the compressed file): This program is an Amiga CLI-based patch librarian for Oberheim Xpander and Matrix-12 patches. You have to use the viewer programme to decompress it for use.

The following discoveries are found in Music Machine. If you have any other tricks and tips in using Matrix-12, please email to me. I will put them here ASAP. Thanks!

From Wed Apr 16 09:45:07 1997
by (8.8.4/8.8.4) with ESMTP
id JAA03587 for <>; Wed, 16 Apr 1997 09:45:05 -0700 (PDT)
16 Apr 97 12:42:05 -0500
16 Apr 97 06:06:47 -0500
Date: Wed, 16 Apr 1997 05:53:40 -0400
From: Jens Johansson <>
Subject: New to list... I'll jump in.

Here are some weird M12/Xp programming discoveries. Hope any of these aren't old news. I made them up all myself and wrote them down many years ago on a
piece of paper I just unearthed! Some of these may be stupid ways of doing things..

* How to get non-12-tone scales:

Variant 1:
OSC page2 controlled by KBD. matrix mod: KBD + 62 and KBD + 49 = 19 steps/oct

KBD -> TRK1 -> LAG; OSC page2: controlled by LAG only.
TRK1 values 0, 15, 31, 47, 63 = 6 steps/oct (whole-tone scale)
TRK1 values 47, 39, 31, 23, 15 = -12 steps/och ("reverse" scale)
(Variant 2 has the benefit of easy routing to the filter as well.)

In general, the "VCO page2" hardwire LAG is a good path to weirdness. Anything can be plugged into the LAG input and be routed straight into the brain of the VCOs for much more extreme frequency sweeps (and speed?) that you'd get from patching via matrix modulations to VCOXFRQ.

* RAMPs sound like they have faster attacks than ENVs. (Or is this my imagination?)

* You can obtain in-phase frequency multiples of of a LFO signal (triangle wave) by routing it thru a TRK set to [0, 31, 63, 31, 0] (this doubles the frequency). [0, 63, 0, 63, 0] gives a crude quadruple-frequency wave. This process can be repeated with more TRKs in series of course.. (if it weren't for the fact that the whole matrix mod scheme implemented in software, there could be some interesting FM effects to be had there!)

* The TRKs can also be used to create weird LFO wave forms. Or to skew the distribution of LFO noise, or LFO sample-hold-noise. Or to create weird envelopes for that matter, by routing a RAMP into one.

* Another good use of the TRKs are to assign a bipolar lever (or anything else) to more than one parameter: two TRKs with values such as [0, 15, 31, 31, 31] and [31, 31, 31, 47, 63] split up the lever into one only-negative and one only-positive signal.

* I managed to get a pretty good simulation of "triple BBD" type ensemble effects circuit (a la the CS80 or PolySix or some stringers) using six LFOs and three VCOs per voice. I don't remember exactly now how it went but it used key trigging of groups of three LFOs that had the same frequency, but different phases. It used two different single patches, one with four LFOs/two VCOs and one with two LFOs/one VCO. The LFOs modulated the VCO frequencies. This may sound like no big deal at all, but it really gave a nice warm pad with a very "tweakable" lushness parameter, that had the added bonus of not having to use PWM.

* The LAG unit can be used for sample-and-hold. Set to EXPO, RATE=63. Modulate the rate with the "GATE": GATE=63 => sample, GATE=0 =>hold.

* You can derivate a signal crudely by patching the "raw" signal as positive modulation and a LAGged signal as the same, but negative, modulation. (to the same destination then of course.) [Also adding positive modulation, and negative quantized modulation, of the same source to the same dest is

* The SAMPLE wave of the LFO can be used for a "multiply" type hack. Set the frequency to 63, plug one factor into the sample input, and modulate the VOL
of the LFO with the other factor. The LFO output is a rough appx of the product. (It's a bit slow though, even at RATE=63) I've for instance used this to set up a "key-dependent pitch change" to a lever..

* Also a LFO can be used to S/H with a bit more hassle if the LAG gen is used for something else; here's an example that enables you to sample a signal with PED2 (63=SAMPLE, 0=HOLD). LFO1 OUT has the sampled signal.

[RAMP1/RATE=0/EXTTRIG] -> sample in of [LFO1/WAVE=SAMPLE/RATE=63/VOL=0]
[LFO1] -> TRK1 [14,29,33,47,63] -> sample in of
[sample/hold input] -> sample in of [LFO3/WAVE=SAMPLE/RATE=63/VOL=0]

Matrix modulations:

* Here's a really weird one (wastes a lot of LFOs and modulations): a flip-flop. PED2 will toggle the output.. and the "output" of the flip-flop

[RAMP1/RATE=0/EXTTRIG] -> sample in of [LFO1/WAVE=SAMPLE/RATE=63/VOL=0]
[LFO1] -> sample in of [LFO2/WAVE=SAMPLE/RATE=63/VOL=63]
[LFO2] -> TRACK1 [63,63,63,31,0] -> sample in of

Matrix mods:

[ Hey I wonder if you could construct a counter out of these flip-flops, if there were more LFOs?? :) ]

Some other random stuff:

* It's a good idea to keep a non-sounding patch in S00, so you can very quickly key in '00' in voices you temporarily want to mute in Multi mode...

* Hitting '99' saves you a few milliseconds of time if you want to just enter the maximum value of '63'

* I opened the pitch block up on my M12 and turned the levers around, that way they correspond physically with most other synths. The polarity then needs to be reset via software, though, but at least you don't have to change every patch if you like the pitch wheel on the left and the mod wheel on the right, like me!

Possible M12 bugs:

* The VCO's are a bit unlinear in the extreme low end.
* PRS leaks over into PED2 and PED1.
* High lag rate and positive matrix modulation result in lower lag rate.
* LFO rates depend slightly on the number of modulations??
* I suspect the possibility of "patch viruses" that crash the synth
inexplicable until you zero out the RAM completely.

Jens. (offline) <> (

The following tips and tricks are from Xpansions ! The Newsletter of the Xpander Users' Group Compilation, Issues 1-7 1986-1988 and also Issue 8 Summer Solstice 1991:

by Dan Barrett

Although the Xpander and Matrix 12 are not hard-wired to do tricks like dynamic panning, Rhodes Chroma-style "steel guitar" pitch bending, and keyboard inversion,these are all relatively easy to accomplish. The solution lies in creating two or more nearly identical Single Patches and combining them creatively as Multi Patches.

Dynamic Panning
How would you like a sound that sweeps from speaker to speaker for a stereo tremolo? Or how about a sound that moves through space depending on how hard you play? Here is the basic idea:

1. Create a patch you like. Store it in TWO locations (for instance, Single Patches 10 and 11). We'll call them patch A and patch B.

2. Make a change in the two patches, so that you have something OPPOSITE happening in each patch. For instance, you can have patch A become LOUDER as you play harder (modulating a VCA with velocity) and patch B become SOFTER as you play harder (this is done by negative modulation of the corresponding VCA by velocity).

3. Go into Multi Patch mode and layer patch A on top of patch B (NOTE: this is easily done by selecting three voices per patch, assigning them to two different Zones, and making the Zones cover the same notes on the keyboard).

4. On the PAN Multi Patch page, pan all of the voices playing patch A fully left, and all of the voices playing patch B fully right. That's the basic idea. Now, soft playing is only heard in the right speaker, and hard playing in the left! What's really happening is that certain voices are becoming louder and softer. Because the two patches sound identical, however, it sounds like ONE patch moving from speaker to speaker. Once you understand this basic idea, you can do countless variations on the theme:

Use an LFO to sweep the sound from speaker to speaker. In both patches, have LF0 1 modulate a VCA from silence to full volume. Also, set RETRIG to SINGLE on LFO 1 Page 2. Now, set the LFO 1 RETRIG on Page 1 in patch A to 0, and in patch B to 31. Layer the sounds as above, and presto! Stereo panning! *You can use an envelope generator instead of the LFO, for a single, controlled sweep between speakers (and perhaps Tracking Generators and Ramps for other effects).

There is no reason why the two patches have to sound the same; experiment!

For an all-out psychedelic experience, create 6 (Xpander) or 12 (Matrix) different patches with different levels of positive/ negative velocity sensitivity. Layer them all on top of each other, pan them all differently and you have a massive, combined instrument whose "panning" changes at your command. If this doesn't impress your ears, you may as well trade in your instrument for a Juno-106.

"SteelGuitar" Pitch Bending
The Rhodes Chroma Polaris allows you to do the following: hold down a chord, press the sustain pedal, then release some notes in the chord, and then move the pitch wheel. Only those notes that are STILL HELD ON THE KEYB0ARD will bend! Perfect for those "steel guitar" licks, but the Xpander and Matrix aren't set up to do this ... or are they?

1. Create Patches A and B identically, as above.

2. On patch A, set :
LEVI on Page 2 of VC01, VC02, and VCF to ON (underlined); ENV 1 to 0 0 0 63 0 63, modulating VCF FREQ, ENV2 to 0 0 0 63 33 63, modulating VCA2.

3. On patch B, set :
LEV1 on Page 2 of VCOI, VC02, and VCF to OFF (no underline); ENVL to 0 0 0 63 0 0, modulating VCF and VCA2, where its RELEASE is modulated +63 by PED2 and its AMP is modulated -63 by ENV2; ENV2 to 0 0 0 63 7 63.4.

4. Notice that patch A has full sustain but quick release. Patch B, without holding down the pedal, has the fastest possible envelope. However, if you hold down the pedal, it has RELEASE ONLY.

5. Create a Multi Patch with patch A layered onto patch B.

Now, when you play normally, only patch A will sound. If you hold down the pedal, patch A will sound on the attack, decay, and sustain, but will die on the release; patch B, however, will kick in on the release. Since patch A bends with LEVI and patch B doesn't, only those keys HELD DOWN when you depress the pedal will bend!

Inverted Keyboard
Here's my trick for the "inverted keyboard:" it's quite simple.

1. To learn this mod, initialize a patch by holding STORE and hitting CLEAR. (Later, you can apply this method to any patch you like, but this way is easiest.)

2. Then turn VCO2's volume too, to simplify things at first.

3. On Page 2 of VC01, turn off KEYBD so it's not underlined. This makes VCO 1 not modulated by the Keyboard. If you play the keyboard, all the notes will have the same low pitch. Back on VCO 1 Page 1, set FREQ to 63.

4. Goto theVCO 1 FREQ modulation page, and choose KEYBD as your mod source.

5. Modulate negatively with the KEYBD (as many times as necessary) until you like the keyboard tuning- it will be inverted. It IS possible to modulate enough so that you get a standard 13-key octave, inverted. If you have problems tuning to an exact range, hit two notes (say, 4 octaves apart) several times each so all of your voices are playing one or the other, and turn both VCAs up so the notes are droning. Then fine-tune the modulation until you hear the right number of octaves between the two notes, and turn the VCAs off again.

6. Now you're done, unless the resulting keyboard range is still too low for you. If so, make a Multi Patch from your Single Patch, and TRANSPOSE up two octaves!
NOTE: you might want to change the Page 1 value of VCO 1 FREQ so a chosen note matches what it would be on a standard keyboard- say, middle C, for instance.

by Dan Barrett

The standard DADSR envelopes on the Xpander and Matrix are LINEAR in function; that is, the Attack, Decay, etc., each occur at some constant speed from beginning to end. However, it is possible to make the envelopes EXPONENTIAL or LOGARITHMIC in function, i.e. have envelope segments that "start slow and then speed up" (exponential) or"start out quickly and slow down" (logarithmic). Why do this? It makes for some interesting envelopes, that's why! I believe that certain natural sounds are "exponential," as some of my patches sound more realistic when I use exponential envelopes.

The trick? Modulate the envelope segment (ATTACK or whatever) with the ENVELOPE ITSELF! For example, if you modulate theATTACK of ENV 1 with ENV 1 positively, you'll get an exponential attack, and if you modulate negatively you'll get a logarithmic one. I have found this trick useful for sounds with a breathy attack (brass, woodwind, human whistle), cartoon effects, and what I'm forced to call "original sounds"because I can't describe them in print! The amount of modulation is CRITICAL: too little and you won't hear the effect, and too much and it will seem as if the envelope is taking forever.

The best way to hear the effect "directly" is if the envelope is modulating oscillator pitch, but it has many other uses as well. By the way, you can do this trick with other "modulatable modulation pages." Modulating an LFO's SPEED or AMP with itself is wild- it changes the shape of the LFO wave!

by Dan Barrett

Here's a fun patch idea. Create a typical 2-VCO synth sound, then use an Envelope to modulate:

1. Both VCO amplitudes, from 63 to 0;

2. VCF Resonance, from 0 to 63;

3. FM Amplitude, from 0 to 63 (try setting FM Dest. to VCF).

This set of modulations should "pan" your sound from a standard 2-VCO sound to a total FM patch. In place of an envelope, you could try this with a Ramp, a Pedal or Lever, an LFO,or whatever. Another neat trick, setup two identical Envelopes for any sound, one Single triggered, one Multi-triggered. Set one so its Amp is at 0 and the other at 63. Then use a Pedal to vary the mix of these two types of envelopes, so you can switch from single to multiple triggering within the same patch.

by Dan Barrett

Matrix-12 Owner's Manual Errors:
A. Pages 20 and 109: NOISE waveform missing from diagram.

B. Page51: It is false that the Matrix Service Manual has "more information". The info is actually in the Xpander Service Manual.

C. Page 61: MIDI Tune Request. The manual says that the Matrix will ignore this request. However, the Xpander/Matrix-12 MIDI Document says that the Matrix will
respond to this request. The manual is correct, according to Oberheim (but my Xpander responds to Tune Requests).

D. Page 134: The statement that the SPEED control has no effect for these two waveforms is false-the SPEED control does affect the RANDOM waveform.

E. Page 137: The paragraph labelled "Stopping an Envelope" should be on page 132, just before the LFO X section, not here.

Matrix-12 Owner's Manual for Xpander Owners: I recommend that Xpander owners that are having a hard time programming their instrument should get the Matrix 12
Owner's Manual. It is around 200 pages long and full of examples, and the reader may ignore references to features exclusive to the Matrix (such as the keyboard).

Undocumented Feature on Tune Page:
You can abort a TUNE ALL in the middle of its operation (Don't you hate it when you hit the TUNE ALL button accidentally during a gig and have to wait for it to finish??) To abort, just press the ALL button again.

The TUNE will abort after it finishes with whatever section (VCF, RES, etc.) it was on when you gave the abort command. (When I tried this, it would actually stop the tune in the middle of a section. Also, call it a bug or a useful feature, your choice, while the MIDI MUTE button on MASTER Page will cause stuck notes to immediately go to the RELEASE segments of their envelopes, hitting the TUNE PAGE button will instantly mute the instrument.)

Software/ Hardware Bugs:
(Matrix 12)
If you are using a quantized modulation and you go to the MOD ROUTINGS page to look at it, the quantization will cease to function. It will return if you jump to another patch and back.

(Matrix 12/ Xpander)
If you initialize voice (hold STORE and hit CLEAR) while on a Mod page, you'll get the OBERHEIM patch with an unwanted RAMP 1 modulation on that page. That's because RAMP X and CLEAR share the same, button.

(Matrix 12)
External Triggering works ONLY for the set of six voices currently showing on the display, even in Single Patch mode. The other six won't trigger unless you hit the BANK SELECT button.

(Matrix 12)
There is a bug in the pressure sensor: hard pressure on the keyboard, even if you're not using PRESSURE modulation in a patch, will activate modulations supposedly under the control of PED 1 or PED 2. PRESSURE is "bleeding through" into PEDAL modulation! (NOTE: Dan states that all of these bugs have been reproduced on other instruments than his own, and that Oberheim has informed him that the first three problems will be fixed in the next software update and the last is being investigated.)

by Mike Metlay

(The following article was originally written for Electronic Musician magazine as a discussion of how to hook the Xpander tip to MIDI and CV external devices effectively, and has been extensively written and revised (in other words, cut to pieces) for the Xpansions! audience. In future issues, I may present more concrete tips in this area, but for now I want to introduce the concepts themselves. WARNING: this article contains pre-MIDI nostalgia in LARGE gobs, and may make hard-core new toy faddists ill. You have been warned!)

No other synthesizer makes use of external control devices like the Xpander. Its combined MIDI / CV flexibility lets the user approach the process of sound creation in a vast number of weird and wonderful ways. This is a pet subject of mine, so I'd like to give some tips on how to "rediscover" the marvelous things Control Voltages (CVs) used to do before MIDI came along, as well as how to use MIDI to open up new and bizarre vistas of sound control.

What We Have To Work With
The Xpander can accept six sets of CV and Gate inputs, a single trigger input, two external voltage sources called "Pedals," and the following MIDI information: Note-on and Note-off, Velocity, Release Velocity, Channel Pressure, and up to five MIDI Controllers. Each of these may be applied to the Xpander's pages as modulation sources. The remaining modulation sources are Tracking, Lag, Vibrato, LFO, Envelope, and Ramp. Of these, we will not concern ourselves much with the latter four, as they're internal to the architecture; we only note that they may be triggered from the external Trigger input, letting us start envelopes or sync LFOs externally. We'll glance at Tracking and Lag later. This leaves us with eight externally-activated modulators; of these, Keyboard and the two Pedals may be run by rear- panel input voltages. On the MIDI side of things, the two Levers may be set to respond to any pair of Controller numbers desired. This also applies to the two Pedals, if the user desires to operate them from MIDI rather than from external devices plugged into the Pedal jacks, and to Pressure modulation, which can be set to respond to a MIDI Controller if the user's keyboard isn't pressure sensitive.

Note that the Levers and Pressure are only accessible via MIDI, the Velocity modulator can respond only to actual MIDI Velocity data, and thus is useless if the user's keyboard is not velocity-sensitive, and a similar (and far more common) restriction applies to Release Velocity. But will we let these petty limitations stop us from wreaking audio havoc with our Xpanders? Of course not! After all, the Xpander is what one makes of it, perhaps more so than any other synthesizer on the market; it was originally designed to be operated from previously-owned MIDI or CV devices, and will do AMAZING things for surprisingly little if you're willing to invest in some dinosaur technology or some cheap MIDI doodads....

"But I've Already Got A Keyboard!"
I know that. This article is intended to get you (only temporarily, if you insist) out of normal "keyboard and two wheels" modes of thought, if all you want to do is noodle on the keys, and dial in an occasional pitch bend or vibrato, that's fine, but you can do that on any synth. Still, we do have them lying around, so let's take a moment to discuss ugh-keyboards.

The defaults for the two Levers are MIDI Pitch Bender and Controller 1; by convention, most MIDI synthesizers that have two left-hand devices assign them to these - two Controller numbers, and thus allow the Xpander to be operated directly from the outboard synthesizer. Velocity and pressure sensitivity are understood automatically by the Xpander. If you own a synthesizer with preset MIDI Controller numbers for other devices, such as damper or volume pedals, a portamento switch, or a breath controller, any of these can be assigned at will to be Pedals or Levers. What if you don't have such things? Not even a damper pedal? Fear not; we'll get to helping you in a moment
If you're using a voiceless MIDI master keyboard to operate the Xpander, you probably know by now that most of these keyboards allow polytimbral splits and layering as well as assignable control devices, but that these functions are duplicated by the Xpander.

As you add more traditionally-designed synthesizers to your setup, they will probably be limited as to which MIDI Controller numbers they can receive, and the Xpander's input flexibility becomes a boon, you'll be free to pick and choose which, if any, of the devices on your master keyboard control the Xpander along with the more conventional synths. For instance, if you want every synth in your MIDI network except the Xpander to follow pitch bends, it's easy to set Lever 1 on the Xpander to read something other than Pitch Bender data.

MIDI Controls (for the Xpander or Matrix 12)
Suppose, despite the obvious advantages displayed by these CV devices, you're still too paranoid to risk cash on old technology. Or suppose you own a Matrix 12, with which CVs aren't an option, and you've tied up both Pedal jacks with stuff I've just talked about before getting this far. What then? Why, go all MIDI, of course! Even without the Trigger and CV/Gate jacks, your options are far from limited ... just pretend that you have a modular synth with one big disadvantage (you have only five knobs to twiddle at a time, rather than all of them) and one big advantage (it's PROGRAMMABLE!). There's an easy way to take this idea to its full extent with only a small investment in cash terms. It's called a MIDI Controller box, and its basic function is to convert the use of various devices into MIDI data. J L Cooper makes a very powerful one called the Expression Plus, but I personally prefer the Yamaha MCS2. It was originally designed to give full MIDI control of synth stacks to a person playing a MIDIed piano, and is essentially a small MIDI merge box that adds its own output to the data coming from its input keyboard. This output includes data from its built-in pitch and modulation wheels, sliders, switches, buttons, foot pedals, and even a breath controller, all in various programmable configurations. This lets you choose which control devices you'd like to use, and which ones are used for which patch. (It may also be possible, if the voltage ranges are compatible, to patch old CV devices into the foot pedal inputs of the MACS.) If you're serious about giving this sort of control a try, it would be an excellent investment ... and it has a number of other uses around the MIDI studio as well, like controlling MIDI effects stacks, changing programs on the fly, etc.

NOW We Go Bananas!
Once you've chosen your setup and gotten used to running it smoothly, it's time to really put the Xpander through its paces. Because the Xpander can use input from MIDI or CV devices in dozens of different ways, the number of electronic Xotica you can add to your performances grows to the point of ridiculousness. To heck with kid stuff like vibrato or volume control; there's serious fun to be had here! Consider the lowly VCA; it's just a loudness control, right?

Not if you're sly enough to try a stunt like the one I pulled on a friend of mine in the studio recently. At the time, his DX7 had an FC3A footpedal controlling its volume, which I enabled on my Xpander as a MIDI source for PED 1 after MIDIing the two synths together. He was expecting me to set up direct VCA control via the pedal to make it a nice, ordinary master volume pedal for both synths; instead, I inverted the VCA modulation on the Xpander, so the further down he pushed the pedal, the softer the Xpander got-and he found himself in full control of a smooth, even DX7/Xpander crossfade, without even having to touch the mixer!

Not weird enough for you? We haven't even begun to use Tracking or Lag yet! Since the Tracking Generators can convert a linearly changing signal to a nonlinear curve, we may alter the response of our devices to our tastes. Imagine a volume device that's loudest in the middle of it's throw, so that swells turn into fades with one smooth push or pull of the relevant knob. The Lag Processor, which smooths out sudden transitions, can turn an ordinary switch into an "automatic fader" with many practical uses. One of my favorites is an improvement on the "slow mod button" feature found on the Roland Axis; we patch a switch or a sustain pedal into the Lag processor, and modulate our sound via Lag. Then, when we press the switch, the modulation gradually fades itself into our sound, rather than turning on instantly at full depth.

By now I think you should have the idea. Go wild! If you haven't got enough knobs, or enough hands, to perform all of this weirdness by yourself, remember that you have the option of up to six different pitch-control sources for your Xpander at any time. You can play a MIDI keyboard and have a CV sequencer doing things at the same time, and still have room to bring your old set of bass pedals into the digital age, too. Add to this the MIDI gizmos that will work on the Matrix 12 and the many signal sources patchable through the Pedal inputs, from analog sequencers to alpha wave headsets, and you may never look back! If you're feeling up to a challenge, try programming and using your Xpander with voltage or MIDI controllers alone ... and no keyboard at all. The results might surprise you.

by Lionel Cassin

Five envelope generators! Five LFO'S! Three tracking generators! However, after using the Xpander for about a month, most synthesists' amazement is replaced by dismay: "how am I going to utilize all these components?" So I singled out the envelope generators and came up with the patch I've contributed to this issue's Patch Page, which I call TDRIPOFF.

The key to TDRIPOFF, and any delay simulation patch with multiple taps and adjustable delay time, is the delay segment of the Xp's envelopes. It doesn't take long to adjust the delay segments of the 5 ENVs so that they fire at regular intervals ( 1 original and 4 taps), but without adjustable delay time, this patch's usefulness is limited. The next step is to use some continuous controller (Xp footpedal, Continuous Controller slider on Xk, Data Slider or Mod Wheel onYamaha DX7) and, via the MIDICNTRL segment of the Master Page, make it modulate the delay segments (except the first ENV, which is your initial note). As you increase your controller amount, each of the delay segments will increase in proportion.

This patch will come in handy with computer-based multitracking where efficient use of outboard effects is of paramount importance. Also, after a year of using digital units with incremental adjust buttons, having a big slider under my hand felt good. The speed with which you can change the settings makes TDRIPOFF a good patch for experimenting. Take a look at the data on the Patch Page, which will reveal more interesting features.

Patches of this type allow a wealth of other options: No reason why the EG's have to be fidentical in duration or amplitude. Unlike any other delay line I know of, the third repetition can be louder than the original, while the fourth has a longer sustain. You can get "swing delay" by increasing the PED 2 modulation on ENVs 2 and 4.
Since only the original ENV is modulating VCA2, you have to open the amp a bit to get the other repetitions. Therefore VCA 2 AMP functions as an effect level. It's effective range is between 0 and 27.

You may have to change the PED 2 mod amounts if you use a different controller, but try to find some way around it; finding those numbers is a time intensive task. Make sure your ENVs 2-4 are set to FREERUN, or you won't get a sound.

Stereo delay A la Barrett is self-evident. Create two versions of the patch and do one of the following: cut ENVs 2 and 4 from one, 1, 3 and 5 from the other to get a 5-tap stereo delay, or rescale the second patch to get 10 repetitions. Good luck and send me a card when they let you out of the sanitarium. Use PED 2 to modulate the decays, so that longer delay settings will having a fuller sound and very quick repetitions will be more percussive and less cluttered.
There are quite a number of possibilities. If you stumble across a particularly unusual one, please tell me and the group.

by David Ziegele

Run a RAMP through a Tracking Generator and use the TRACK to modulate a destination such as VCO FREQ, VCF FREQ or RES, or VCO VOL. This is not quite as tricky as using an envelope as an input to the TRACK, but can still result in "CZ" style envelopes or an infinite variety of custom ramps or curves. My patch "Whip 5th" (see this month's Patch Page) uses mod on both VCO frequencies.

Terry Darakis asked last issue for brass patches with LFO growl. My method for this: Set an LFO to SPEED 63, TRIANGLE, AMP 63. Negatively modulate the LFO AMP with a very fast RAMP. Use the LFO to modulate one or both VCO FREQs and/or VCFFREQ. Set the RAMP RATE to a setting that lets you hear the growl for as long as you like before it fades out. One could also modulate the LFO AMP with Velocity, so that there will be more growl at high key velocities, and the growl will last longer. To augment the growl, one can slightly modulate one or both VCO FREQs with a very fast ENV. This gives a quick rise (or fall, if you prefer) in pitch that simulates the spit of a true brass attack. This effect can also be made velocity dependent. My patch "Xpndbrss" (see the Patch Page) illustrates both effects.

The new E! board for the Yamaha DX7 has a feature called "random detune" which, as I understand it, slightly changes each notes tuning in a random fashion, supposedly to give acoustic patches an impression of "realism." Here's a way to get random detuning effect on the Xp: set an LFO to a slow SPEED (say around 10) and a RANDOM waveform. Use this LFO to slightly (or drastically) modulate one or both VCO FREQS. The pitch will be different every time you strike a key. If the LFO SPEED is slow enough, the pitch shouldn't change in the middle of the note. By modulating one VCO FREQ positively and the other negatively, you could get a random chorus effect. Be careful using this mod with FM patches, as they'll be quite sensitive to VCO FREQ variations. (The mod could be applied to the VCF for random timbre changes as well.) My patch "Thumbpno" (see the Patch Page) uses this mod. (NOTE: Dave provided the free patches mentioned above as accompaniments to his tips. They were not taken from the set of sounds he's currently selling.)

by Mike Metlay

Dan's inverted keyboard trick came in just as I was finishing work on an article explaining mictrotonal scales for the Xpander. His idea and mine turned out to be
identical except in modulation sign (negatives vs. positive), So I'll save space by adding my article to his.

To create tunings with more than twelve notes per octave, take any patch you like and try the following modifications: follow the steps Dan has outlined above, but modulate the VCOs with the KEYBOARD positively until you have the right number of keys in an octave. Also, VCO FREQ should be set higher than usual, but not necessarily all the way up to 63.

To get you started, here are some "common" microtonal scalings I've worked out:
Notes Per Octave VCO Freq Mod By Keyboard
31 +63, +59
19 +63, +62, +60
24 (quarter-tone scale) +63, +62, +44
48 (eighth-tone scale) +62, +46

Why the mod settings of 62 rather than 63 ? Because the KEYBOARD modulation isn't a perfectly linear function! There's much less resolution at 63 than lower down, and setting a modulation to 62 gives you more flexibility when fine-tuning a sound. Try it!

(NOTE: The use of microtonal scalings isn't necessarily for everyone. Even with my bizarre approach to music, I rarely use this stuff, it hurts my brain. However, recent articles on the subject in the major magazines, the release of tunable instruments like the DX7II FD, and albums like Wendy Carlos' Beauty in the Beast, have spurred people's interest, and hopefully these hints will let you explore this ground a bit on your own.)

by Ted Greenwald

I don't know- maybe I really want a DX7, but I've been having a lot more fun with my Xpander since I started getting heavily into the FM aspect of the machine.
There's an abundance of powerful waveforms lurking in the FM ratios, and they make life much more interesting than your average sawtooth, triangle, and pulse waves.
Aside from the really noisy ones (which are fun too, and which can usually be made less noisy by fine-tuning one of the VCOs), there are great guitar (fuzz and otherwise), clavinet, reed and bell timbres in there. Also, many ratios react beautifully to the high pass filter modes. Unfortunately, the Xp isn't the ideal machine for FM- analog oscillators are the problem, according to Jim Letts at Oberheim. They don't maintain their tuning stability to the fine degree necessary for FM, and as most of us have found out already, Xp FM can be pretty unwieldy. But all of my best patches lately have come from this aspect of the instrument, so I'd like to encourage more FM programming by offering the following observations about FM and the single Xpander:

1. When you're using FM, always tune your VCOS! Otherwise everything will sound like (editor chickens out at this point and substitutes the word ... ) poop - With many ratios, I find that I have to tune several times during the first hour or two the Xp is on before the pitch stabilizes. Even then, one voice may end up a little bit off after tuning, and must be brought into line by re-tuning again and again until it settles down. This makes mondo FM patches a bit impractical for live performances, unfortunately, but in general, if you can sneak in a tuning, do it whenever possible.

2. With very high VCO frequencies (which is where most of the interesting timbres hangout, unfortunately) the pitch consistency among voices tends to run amok. Some felicitous ratios exist in the high ranges (on my Xp, at least), but by and large I've found the safe upper limit for either VCO FREQ to be 35 or so. Above that, some voices will almost inevitably be out of tune. You may or may not hear it easily- that depends on the particular wave you're making- but it's bound to cause problems eventually.

3. A slight portamento can be heard when a voice goes from a low note to a high note. This is incredibly annoying, but there's no way around it. The best solution I've found is to create a Multi Patch and assign, say, two voices to the lower keyboard half and four to the upper half to limit each voice's possible pitch changes.

4. The following is a bunch of information adapted from the DX7 programming chapter of Steve DeFuria's Secrets of Analog and Digital Synthesis. It'll be useful to DX owners as well as FM- happy Xp users, but I've translated the DX-specific stuff to be more directly applicable to everybody's favorite synth (after the ARP 2600). Some of the information doesn't quite translate, unless you map the DX's harmonic-series-based numbers onto the Xp's chromatic-scale/ fine-tuning numbers. Beyond the first five harmonics, I've left that to you. DeFuria's Analogies are convenient but his Rules are serious data-an initiation into FM arcana that I've never seen printed anywhere outside his book. Make good use of them ... or just program by trial and error. As long as you make good sounds, right?

(NOTE: remember that on the Xp, VCO 1 or the VCF is the Carrier, and VCO 2 is the Modulator)


Carrier frequency will be the most prominent partial. Altering carrier frequency thus determines which partial will be loudest. (This is analogous to resonant filter cutoff in analog synthesis.) Carrier output level determines overall loudness. (Analogous to VCA.)

Frequency of modulator determines overall partial structure, or waveform. (Analogous to choice of waveform.) Modulator amplitude determines strength and number of partials, not necessarily in harmonic series order. (Analogous to altering VCF cutoff.) Rules (we assume the fundamental will be represented by a VCO FREQ of 12): If the carrier is tuned to a pitch in the harmonic series, and... If the modulator is 12,24, 31, or 36 (the first four pitches in the harmonic series), a fundamental will always be present, even if the carrier does not represent the fundamental.

... If the modulator is 12 (fundamental), all harmonics will be present (as in a sawtooth wave). ... If the modulator is 24 (first harmonic), only odd harmonics will be present (a square wave). ... If the modulator is 3l (second harmonic), and the carrier is tuned to a pitch in the harmonic series, every third harmonic will be missing (as in a pulse wave with a 33% or 66% duty cycle). If the modulator is tuned to an odd- numbered harmonic (24,36,43 and so forth), only odd harmonics will be present.

... If the modulator is 4O (fourth harmonic) or higher, some combination of harmonic partials will be present.

... If the modulator is not tuned to a pitch in the harmonic series, some in harmonic partials will be present.

Detuning the modulator will create out-of-tune harmonics, a phenomenon that does occur in nature. The partial produced by the carrier (not necessarily the fundamental) will retain in tune unless detuned as well. A carrier tuned to a low, fixed frequency will, when modulated, produce a tremolo at twice the carrier's frequency.

5. Finally, an FM patch (see the Patch Page). It's the best Clavinet simulation I've come across (if I do say so myself although it's still not perfect. I have no doubts that a DX7 could do better, but I haven't gotten my hands on one to try yet. Using David Ziegele's random detuning tip from last issue would probably help simulate the clavinet's tendency to change pitch slightly from keystroke to keystroke, and also its proclivity for going out of tune. As with all FM sounds, don't forget to tune the VCOs a few times during the first hour or so. Enjoy.

Vladimir Vooss has suggested via letter and cassette that the Xp and Yamaha's FM synthesizers have different tunings. Both of them stray from Vladimir's ideal: stretch tuning. The TX802 is slightly flat, as is the Xp, but the Xp has places where it's sharp. When the two instruments are combined, this discrepancy results in a "sour sound," rather than the grandiose tones we expected from the analog/FM marriage. Vladimir has customized the tuning on his 802 to match the quirky Xpander.

Lionel Cassin: The UNTUNE function may be a boon to industrial noise fanatics who don't want to waste modulation routings to generate random pitches. Goto MASTER PAGE, Page 2, SERVICE, and UNTUNE. What a mess.

by Glenn Workman

1 - Here's a quick method to load several patches off a single cassette. First make sure you've saved your current memory data. Load in the cassette that has the needed patches. Create a Multi Patch that uses the singles that you need. Make sure that that Multi is the currently selected patch. Reload your original data. Your "new" Multi is still in the editing buffer. Now store each "new" Single without leaving the Multi Patch (this is described on page 18 in the Xp manual, page 167 in the M12 manual).

Why this is: The Multi and Single editing buffers are unaffected during a cassette load. The contain complete data for all selected Singles. This means up to seven patches on the Xp (six in the Multi buffer, one in the Single buffer) or thirteen on the M12 (twelve in the Multi buffer, one in the Single buffer) can be recalled after a cassette load. This will also work when saving and loading through MIDI.

Additional Tips: Predetermine where you will put all these new patches. Make sure the currently selected patches in Single and Multi are the ones you want before you reload. Do not move away from these patches until you have stored them or all will be lost. (You may toggle the Master Edit Single and Multi buttons without harm.)

3. I've found a way to coax an extra lever or two out of the Xp/ M12 that can be a big help in a performance situation. For example, if you would like two different vibrato speeds, set up a basic patch with vibrato underlined on page 2 of VC01 and VC02. Now select FREQ for each VCO and add LF01 / +63 as a modulation. Go to LF01 page and set up SPEED 55, square, and AMP 0. Select AMP and add LEV2/-63. When you push LEV 2 one way you get your original vibrato, push the other way and you get your alternate LFO 1 vibrato trill. This requires several things to work. Your instrument should have centered levers likethe M12/M6/Xk so it can be moved positive/negative (some wheels will not do this). It seems most useful on modulations like LFO/ENV amplitude or SPEED settings set near zero. That way the reverse action will not affect it (though that can be a useful crossfade between effects). Variations: This will also work on LEV 1 but it seems best to turn bend off. You can use an on/off pedal to switch between effects. That way it won't just add something but will also defeat something else.

4. Reverse Keyboard: The Improved Easy Method.

Begin with the basic Oberheim patch. On VC01 page 2, turn off KEYBD and turn on LAG. Repeat for VC02 page 2, and VCF page 2. On the FM/LAG page, select TRK l for LAG IN and 0 for LAG RATE. On TRACK 1 page, select KEYBD 47, 39, 31, 23, 15. All done! (This is the DRAWKCAB patch, printed elsewhere). Major chords are now minor, minor chords are now major, diminished chords are still diminished. Now try it on your favorite patch following the same steps. It requires that the LAG be free and that one Tracking Generator is unused. B and F keys are still at their named pitches. If you want to change that, add or subtract the same amount from all TRK 1 values.
Variations: Try TRK 1 values of 47, 39, 31, 39, 47 for a keyboard that goes both ways. 0, 15, 31, 47, 63 for a whole tone keyboard (the lower octave is out of tune this way; 0, 16,32,48,63 will be the same with the upper octave out of tune). 23, 27, 31, 35, 39 gives a quarter-tone tuning. If you really want to know, 18,26,34,42,50 leaves the pitches where [nothing was typed here]

Allan Tamm mentions that if you're driving your Xp with the Yamaha WX7 Wind Controller, "assign the WX7's pressure data to be Aftertouch (Pressure) so you can use the same patches on the Xp with a keyboard. I've had good success with the volume response of the Xp by assigning Aftertouch as a modulation source to VCA2 (in the VCF/ VCA section) twice at an amount of 63. Introducing vibrato independently of amplitude seems to require a foot pedal, unfortunately. I've had good results with the stock flute and oboe voices; brass seemed less satisfying."

Also in the alternate controller vein, George Tucker writes: "For any guitarists having tracking problems, try setting the main output envelope (generally modulation VCA2) to RESET. A new gate starts things all over. I'm having good luck with this for patches where I need speed."

by Doug Shawe

Here's my favorite tip for quick editing when you're doing a session for people who are, invariably, unfamiliar with the Xpander. Most seasoned engineers know that it's better to change the harmonic /frequency content of a signal at the source rather than EQ it. So you'll be asked to make you patches "brighter." The fastest way to do that on the Xp is to go to the VCF and scroll through the Pole Filters available for that particular patch. Not only does this work 80% of the time, but everyone (including the engineer) thinks you're brilliant!

If that doesn't do it, and tweaking Frequency and Resonance doesn't help, move on to another patch, because time is the imperative in the studio. Nobody wants to stand around for five minutes while you try to adjust a sound to perfection. That brings me to the importance of basic patches to the studio keyboardist. Most people in the studio want to hear the tried and know, and arrangers reinforce this habit by including notes in their charts like "DX7 Rhodes" or "Warm Obie Pad." You of course need a decent working knowledge of synth editing, but basic patches are absolutely essential in a studio situation.

Here's an instructive anecdote from some sessions I did on an Epic Records LP (no names-I still want to work in this town) where I brought in my entire key/rack/computer rig. The engineer asked me for brass and I mixed four TX's, an MKS-80 Super Jupiter, an Emax and my Xp for what I thought was a gorgeous brass sound. "It's out of tune!" said the producer. After double-checking with a strobe tuner, I said it was probably detuning within patches to get a fuller sound. "Get rid of it!" OK. "What's that wobbling!?" Good-bye LFOS. "It's too goddam big!!" Out with the octaves. "Still too big!', Eliminate TX's and Emax. "Faster!" Attack=O. "More sustain!" Reset Decay, Sustain and Release. "THAT'S IT!"' Roll tape- the cheesiest, much unnatural synth sound ever recorded. I am quietly embarrassed, but a few dollars richer.

by David Gilden

The other day, I was using a sequencer to drive the Xpander, with a Yamaha MEP4 MIDI event processor in between the two. Using the MEP4, I sent the same MIDI data twice on the same MIDI channel with a slight delay added on one of the MEP4's "Processors." With this, I achieved a thick kind of slow phasing, cutting the number of available voices in half but providing a fatter tone. On the Matrix, which has its own keyboard, try connecting the MIDI OUT to the Matrix's own MIDI IN to obtain the same effect. In this case, the effect is due to the fact that that the synth's internal microprocessor takes some time to convert data to control voltages.

Did you know that when you SYNC the two VCOs (the switch is on Page 2 of VC02) and turn FM to VC0 1 with an AMP of 63, you create a feedback loop that causes the VCOs to act in a very strange, nonlinear manner? Jim Letts, Oberheim's Chief Engineer, was quite surprised to find this out. If you're tight for memory, the Xpander can give you chords, up to six voices worth for a single note-on command: just setup a Multi Patch to play some or all voices in unison, and set them to chord intervals on the Transpose page! For example, if you want a major triad, set one voice to the root, one voice to the third, and one voice to the fifth, put all three voices in the same MIDI Zone, and set the Zone's Mode to Unison (High, Low, or Last, depending on your preference). You can have a different Multi Patch for each chord structure, and access them with MIDI Patch Change commands. This hint is particularly useful for Commodore 64 users; I use the Songstepper from Moog Music, whose 12-channel monophonic setup makes this a must.

And one last bit of advice, for people who've found themselves caught in the dilemma of having created a new patch and not knowing which memory location to store it to. It's true that you'll lose your edits if you try to audition other patches on the Single Patch Page, but that doesn't keep you from hearing Multi Patches! All you need to do is to get into a Multi Patch, audition various Single Patches within that Multi Patch until you find one you won't mind losing, memorize its number, switch back to Single Patch Page and store your edited patch to the number you've memorized. This is an obvious trick, but I hope it'll save a lot of sweat for those who haven't thought of it themselves.

Using The Keyboard Modes: A Tutorial
by Dan Barrett

A few of our fellow XUG members are having some trouble understanding keyboard modes. Looking at the manuals for both Xpander and Matrix-12, I see that the explanations are very short, and contain few examples. This article is a tutorial to make you familiar with the differences between, and the advantages of, the diverse and powerful Oberheim keyboard modes.

This article assumes you already know what a "zone" is. If you don't, please see page 47 in your Xpander Owner's Manual, or page 87 in the Matrix 12 Owner's Manual. I will use the term "Xpander" to mean "Xpander or Matrix-12." Matrix-12 owners will have to substitute "12" instead of "6" in some of the examples.

First of all... what is a "keyboard mode?" When we select a keyboard mode, we are choosing how the Xpander will behave when we press a key. More specifically, we specify exactly which voice will play when a key is pressed. Now, keyboard modes are not the same thing as zones. A zone allows us to group voices together, to be played in a particular range on the keyboard, or on a particular MIDI channel. However, within each zone we still must choose which voices will play when we press a key. This is where keyboard modes come in!

Let's use an example to illustrate. Suppose that all of your Xpander's voices are assigned to zone 1, stretching the entire length of the keyboard. This is the simplest zone setup on an Xpander. Now press a key on your keyboard. Question: Which voice in your Xpander will sound? Will it be voice number 1, or voice number 5, or perhaps all of the voices? The keyboard mode allows us to choose this.

Here is a concise definition of keyboard mode: it is a method for choosing which voice will sound when a key is pressed in a given zone. If you are a computer programmer, it might be a particular voice that sounds, its dot lights up. Your voices are playing in ROTATION; first voice 1, then voice 2, and so on, up to voice 6, and then back again to voice 1. This is the purpose of ROTATE mode. Now, instead of playing one key all the time, play an ascending scale. As you move from key to key, the sound still changes for every note. No matter what you play, each key triggers the NEXT voice in the rotation from I to 6. Now try playing some chords, and watch the little dots again. Can you figure out what's going on? It's really the same thing-each key triggers the next voice in the rotation. But now, a 3-note chord might play voices 2,3, and 4; the next 3-note chord would play voices 5,6, and 1. For what purposes is ROTATE mode good? You have seen one purpose in the above example: having a different sound play each time you press a key. This difference can be subtle, too. Having 6 versions of the same patch, each slightly different, and then rotating through them, can make your patch seem more natural or life like. But it has other uses as well. For example, you might want to insure that voices will play (release) as long as possible after you let go of a key. In ROTATE mode, if you play voice 3, for example, it will not be reused until all the other voices have played in rotation.

Next, let's look at REASSIGN mode. REASSIGN is very similar to ROTATE, but it has a subtle difference. Go back to your fancy multi-patch and set its zone's keyboard mode to REASSIGN. Now play an ascending scale. Once again, you should get a different sound for each note you play. This is the same behavior as in ROTATE mode. Next, choose any key on the keyboard and play it slowly and repeatedly. What happens? You should hear the same sound every time you press the note! Now choose three keys on the keyboard and play those keys only . Note that each key "retains" its particular voice. Watch the little dots in the PROGRAMMER display to see that this is really happening. What does this prove? REASSIGN mode causes each voice to "remain" at a particular key until it is needed at a different key. This is unlike ROTATE mode, which ALWAYS chooses the next voice in rotation. REASSIGN mode chooses the next voice in rotation unless there is "already" a voice "at" that note. What good is REASSIGN mode?

It is good for realistic piano patches, for one example. When you play a real piano, each key is associated with a different string inside the instrument. So it makes sense that a piano patch on your synth should be associated with a "different" voice. This is much like ROTATE mode. However, on a real piano, when you strike the same note twice, it uses the SAME STRING for both strikes. The second strike will tend to cut off the sound of the first. This is very much like what happens in REASSIGN mode.
For our final polyphonic mode, we look at RESET. Its behavior is quite different from that of ROTATE and REASSIGN. To use this mode, go back to your fancy multi-patch and set its zone's mode to RESET. (Also, it will help if the individual single-patches of your multi-patch are sustained sounds. This is not necessary, but it will help you to understand this exercise.) With one finger, press a key and let it go. Then press a different key and let it go. Press many keys, one at a time, always LETTING GO of the key before you press the next one. What happens? Each key you press triggers the same single-patch. More specifically, voice 1 always plays when you hold down one note. Watch the little dots in your PROGRAMMER display to see that this is true. Next, press AND HOLD one key with your left hand. Voice 1 will sound again. With your right hand, press a second key while still holding the first. The new key will trigger voice 2. Keep holding your first key (left hand), but press and release various other keys with your right hand, one at a time. Each key that you press plays voice 2, while your left hand holds voice 1. Watch the little dots again while you do this, to see what is happening. Now hold two notes down with your left hand, and press a third with your right hand. Voice 3 will sound. What is happening here, in general? RESET mode always causes the LOWEST NUMBERED FREE VOICE to play.

A "free voice" is one that is not being triggered (its little dot is not lit up). Voice do NOT play in rotation. What uses are there for this mode? One possible use is with portamento (glide, or pitch lag). Suppose you set up a patch in which the lag processor is affecting the VCO Frequency, causing your pitches to glide from one to the other. This is the classic "portamento" effect. Play a three-note chord low on the keyboard. Let go, and then play a three note chord high on the keyboard. What happens? In ROTATE or REASSIGN mode, the first chord would play voices 1, 2, and 3. The second chord would play voices 4, 5, and 6. In other words, none of the voices would be the same in the two chords, so no "glide" would occur ! In RESET mode, both chords would use voices 1, 2, and 3, so the sound WOULD glide upward from the low chord to the high chord. Note: ROTATE is the default keyboard mode when your Xpander is in single-patch mode. For this reason, your portamento single-patches may appear not to work! Switch your single-patch keyboard mode to RESET (press MASTER PAGE + ZONE and choose zone 1) if you want RESET to be your single-patch keyboard mode.

Here is a summary of the polyphonic modes. ROTATE causes the voices in a zone to play in strict rotation. REASSIGN also plays the voices in rotation, unless you play a key which already has a voice "at" that key. RESET causes the lowest-numbered free voice to play. Through careful choice of modes, you can tailor the way your multi-patches interact with the keyboard.

Next, let's talk about the monophonic ("UNI") modes. They are called UNI-LOW, UNI-HIGH, and UNI-LAST. In one sense, these monophonic modes all seem the same: they cause all voices in the zone to be played in unison, every time you press a key in that zone. You will experience their differences below.

First, turn down your volume. (This will be louder than the polyphonic examples, since all voices will be playing at once.) Set your zone's mode to be UNI-LOW and play a key. All the voices in the zone will sound at once. Look at the little dots, and you will see that they all light up when you press one key. Set your zone's mode to UNI-HIGH, and then UNI- LAST, and playing notes again. Do you notice any differences? You might not ! I'll explain the differences, and then some applications.

Set your mode to UNI-LOW again. Now press and hold two keys at once. This is a monophonic keyboard mode, so only one key can truly sound your voices. Which key wins? The LOWEST one that you are playing on the keyboard. Hold down six keys and listen: the lowest key is sounding all of the voices. Try the same experiment in UNI-HIGH mode. You will discover that the HIGHEST key is the only one that triggers all the voices. This illustrates the difference between UNI-LOW and UNI-HIGH modes; if there are several. keys held down, the lowest key (in UNI-LOW mode) or the highest key (in UNI-HIGH mode) is the "winner" that triggers the voices. Here is another exercise to feel the difference between UNI-LOW and UNI-HIGH, and it will lead us to understand UNI-LAST mode. In UNI-LOW mode, hold down two keys simultaneously. The low key wins. Continue to hold the lower key, but let go of the higher key. What happens to the sound? Nothing, because your lowest key has not changed. Let go of all keys. Hold down two keys again. Continue to hold the higher key, but let go of the lower key. Surprise ! The voices all jump up to the higher key. Why? Because it is now the lowest key you are holding. Keep holding that higher key, and press the lower key again. The voices jump to the lower key. Still holding the higher key, press and release the lower key repeatedly. Hear the voices jump from key to key, as your "lowest" key changes. Now, try the above experiment again, but in UNI-HIGH mode. You should get the exact opposite results. Releasing the lower key should have no effect, but releasing the upper key should cause the voices to jump to the lower key.

Lets try doing some "hammer on" electric guitar playing in UNI-LOW mode. Set up your multi-patch so all voices are playing the same single patch. Make the sound very sustained even the "voice init patch" (press STORE-CLEAR) will do fine. Let go of all keys, and then hold down a "G" key in the middle of the keyboard. You will hold this key down throughout this exercise. Next, press and release the "F" key immediately below it, as quickly as you can, repeatedly. You should hear a trill. Stop. Now press and release the "D" key immediately below the "D" as quickly as you can, repeatedly. Stop. Finally, still holding the "G", alternatively press AND RELEASE the "F" and the "D", repeatedly, for true "Eddie Van Halen" action. (Make sure you release the keys completely, or you won't get the effect.) Now add some pitch bend...


Um... excuse me. You can use this general technique to "cheat" and play things faster than your fingers could play otherwise.

Try playing the introduction to "Fooling Yourself" from THE GRAND ILLUSION by Styx in UNI- HIGH mode, or parts of Bach's" Toccata & Fugue in D minor" in UNI-LOW mode.

Now, it's time to try using UNI-LAST mode. In this mode, it is not the lowest or highest key that triggers the voices. Instead, it is the LAST key (in time) that you play. Try our 2-key experiment above in UNI-LAST mode. When you release a key, the sound never jumps back to an old key. Only the LAST note you play is significant.

Let's summarize the three monophonic ("UNI") modes. UNI-LOW always causes your lowest key to play the voices in the zone. UNI-HIGH always causes your highest key to play the voices in the zone. UNI-LAST always causes your last key to play all the voices in the zone.

At this point, some people say: "So what? Why do we need all these different monophonic modes?" Well, the true power doesn't become apparent until you try using a monophonic zone layered with a polyphonic zone. Remember, each zone can have a different keyboard mode!

Set up a multi-patch in which voices 1-4 play a sustained sound, and voices 5-6 play a lead synth sound. Assign voices 1-4 to zone 1, in ROTATE mode, and voices 5-6 to zone 2, in UNI-HIGH mode. Set the limits for both zones to be 0 to 127, so they overlap. We now have a monophonic sound layered, over a polyphonic sound. Play a chord, and what do you notice? The monophonic sound always plays the HIGHEST NOTE OF YOUR CHORD! As you play different chords, the lead sound follows your highest key. Instant melody! Now set zone 2 to UNI-LOW and play a chord. The lead sound always plays the lowest note of your chord. Instant bass! Play with both of these setups for a while, and see what you come up with. Be creative! Personally, I use ROTATE layered with UNI-HIGH to play the last section of "Song For America" by Kansas. Through clever fingering, I can play the high melody at the same time as the string chords.

Layering a polyphonic zone with a UNI-LAST zone is a little harder to control, in my opinion. With a little work, one can weave a melody among chords. For a really strange keyboard mode, layer three zones as in the following description. Assign voices 1-2 to zone 1, voices 3-4 to zone 2, and voices 5-6 to zone3. Set the limits on all zones to be 0-127. Make zone I UNI-LOW, zone 2 UNI-HIGH, and zone 3 UNI-LAST. If you play one key, you get all the voices. If you play two keys, you get voices 1-2 on the lower key, voices 3-4 on the higher key, and voices 5-6 on whichever key you played last. If you hold two keys and play others between them, you get only voices 5-6 on the "between" notes.

My goal for this tutorial was to give you experience using the six keyboard modes. There are many different possible ways to use and combine these modes. Different people prefer different modes for particular styles of music or particular kinds of patches. I hope I have motivated you to experiment with keyboard modes. Most other synthesizers give you no choice over keyboard mode: often you're stuck with ROTATE for the entire instrument. Be glad you have a choice on your instrument ... let alone a choice for each zone! So, take this freedom and make the most of it!

How To Test An Xpander
by Mike Metlay

For those of you who might be shopping for a used Oberheim in the future, or who have recently purchased one, here's a little test I wrote up. Maybe you'll find it helpful; the fellow for whom I wrote it certainly did.

This test is in two parts. The first is relatively straightforward, and can be easily performed in any store; the second is more involved, and perhaps is best suited to when you have the instrument set up at home. Where's the dividing line between the two? You turn off the offending voice with the appropriate button or knob. (If any voices didn't light up their LEDs before, you can now see if those voices were disabled here. If they were, turn them on now, and retune the instrument.) Now retune the instrument. If everything else passes, then you may have a bad voice chip. Fixable, but haggle like mad to get a better price: after all, you're now stuck with a 5 voice synth!

Now hit the MASTER PAGE button, then PAGE 2, then the button under the word SERVICE. (If the guy in the store isn't sweating yet, now he will be.) If everything on all the displays suddenly lights up, don't panic! Just hit PAGE 2 again, and you'll see what happened: the DISP utility on the Service page uses the same button as the one you used to select SERVICE, and the button just double-clicked. If that didn't happen, press DISP. Examine the LED displays for bummed-out segments: even one bad one in the wrong place can be a big hassle. When you're happy with what you see, press PAGE 2 to shut it off. (If the display segments all cycle through one by one instead of all coming on and staying on at once, this may indicate obsolete software; see below.) Now press the button under LEDS and watch the red LEDs all fight up, then cycle on and off in sequence, and look for bad ones. Do it again if you're not sure. Then hit the MEM button: it should then read ROM 0 OK, ROM 1 OK, ROM 2 OK, ROM 3 OK and put you back on the SERVICE Page. Hit VMEM: the display should read TESTING for a minute or two (it takes a while) and then read VMEM OK. Press Page 2 to get back to, the SERVICE page, and hit MASTER PAGE again to get out of that (You can hit the UNTUNE button if you want while in SERVICE page; this wrecks VCO tuning for some type of diagnostic. Retuning the instrument should line everything up again. Don't touch the two DAC-related buttons.)

Hit MASTER PAGE and PAGE 2, then select VERSION. The center display should read: MAIN PROCESSOR VERSION 1.2, VOICE VERSION 1.4, CASSETTE VERSION 1.0. Any numbers lower than these drastically lower the value of the Xpander, as its original software had a lot of bugs. (It also tells you that the instrument is a pre-ECC beast; a real find!) The software is upgradable, but in Oberheim's current state (they're retooling their service system) it could take a while to get new ROMS. (If you want to scare the salesman, ask to control the Xpander with a Roland keyboard; more on this in a moment.)

Now, on the MASTER page, press MIDI. You'll be given a list of subpages to choose from: CHANNEL, CTRLS, ENABLES, SEND, RESET, MUTE. Hit CHANNEL. Is the set Basic Channel equal to the master keyboard's? Hit PAGE 2 or MASTER PAGE and MIDI, then select CTRLS. Spin the knobs to set the controllers equal to whatever MIDI data your master is sending. For our example, set LEV 1 to BENDER, LEV 2 to 1, PED 1 to 7, PED 2 to 64, and PRESS to PRESSR. For a Yamaha synth, you might have to use 2 and 4 for the Breath Controller and Data Slider, etc. Use good judgment, and let the salesman help you: after all, you're not buying the master keyboard. Get back to the MASTERPAGE, then hit MIDI and select ENABLES. As you underline the various options or take their underlines away by hitting the appropriate buttons, you should alternately enable and disable the Xpander's ability to receive Patch data, MIDI Controllers, SysEx, etc. Not all of these are easily testable; use your best judgment. The other MIDI options are SEND (don't do this unless the Xpander's MIDI out is hooked to a data storage device like a computer librarian), RESET (don't do this!), and MUTE. Hold a note on the keyboard, and press MUTE; the note should stop.

Okay, now play a few notes and listen. No out-of-tune notes? No weird detuning or phasing problems? All righty. Now press VCO 2, and hold a note while you turn the knob under VOL to zero. One oscillator should fade away, leaving the other. Now press VCO 1, then PAGE 2, and push the button under TRI to underline it: you've added a Triangle wave form. Experiment with turning each wave on and off; make sure they sound okay. Press PAGE 2 again, or VCO 1 again, to get back to Page 1. Vary the PW up and down, with only a pulse wave playing. Make sure that you hear the duty cycle change. Turn the FREQ, DETUNE, and VOL knobs and make sure each one works as advertised. When satisfied, leave the VCO full on, with on a Triangle wave selected.

Repeat this procedure for VCO 2, with VCO 1 turned down. Then listen to both VCOs with triangle waves together: try detuning them and then tuning them and make sure their tunings line up without too much beating when they have identical settings. On VCO 2, listen to the NOISE output, and try SYNC. The two oscillators should hard-sync to one another. The other values on Page 2 control whether the VIB (global oscillator) controls the oscillator, or LEV 1 (your pitch bender), or KEYBOARD, or LAG. Shutting off KEYBOARD means the pitch of that oscillator shouldn't track; shutting off LEV 1 means the bender won't work. Satisfied?

Go to the VCF/VCA page, and try out the various controls (Don't forget Page 2). You'll notice that there's a period after FREQ and VCA 2. That means that they're being modulated. Hit the button under FREQ.

It should say ENV 1 and list a number. Turn the knob to make the number go up and down, and hear the effect on the frequency of the filter cutoff. Do the same for the VCA 2 modulation. (You exit a Modulation page by hitting the VCF/VCA button again.)

Now to test the machine's response to MIDI control. There are two ways to do this: the easy way is to flip through the machine's presets and try out your controllers: velocity, pressure, etc. The hard way is to set up modulation routings in your initial patch; don't waste the salesman's time. You should assure yourself that each controller does something (although the bender may not bend pitch, the sustain pedal may not cause sustain, etc.: they can be set to do anything!). If it seems that one controller isn't doing anything at all, recheck the MIDI CTRLS subpage and make sure the Xpander and master keyboard are talking to each other properly (If PED 1 isn't set to 2, the DX7's Breath Controller won't do anything). This is where a Roland controller will fail if the synth's software is an old version: the sustain pedal won't sustain notes, no matter what, if the controller is a Roland, but it'll work fine for other makes. If all seems well, then you can be pretty sure that there's nothing flaky about your machine. If you want to use the analog PEDAL inputs on the back, set the PEDALs to PEDAL 1 and PEDAL 2 on the MIDI CTRLS page and try them with sustain and volume pedals. Polarities and throws vary a lot, though, so don't be too disappointed if the inputs don't work exactly as suspected. (Yamaha foot controllers won't work at all.) I'll also skip a description of how to test the CV and GATE inputs, the TRIGGER IN, and the chain pedal. You're not dumb.

One last test: plug the amp into the MONO output, and be sure all voices sound with equal volume. Then go to the MASTER PAGE and select PAN. Plug the amp into first the LEFT, then the RIGHT outputs, and turn the knobs to vary the pan position of each voice. Make sure that each voice pans properly, and that there's no bleed
through on the LEFT output when the voice is set to the RIGHT, etc. The pan positions are: LEFT, LF2, LF1, MID, RT1, RT2, RIGHT, DIR. DIR should select the voice's individual output and remove it from the stereo mix. Verify this for each voice.

Now, flip through the presets some more and enjoy the machine. You can check each little part of the architecture when you have a manual and some free time in hand. Have fun!

In the interest of saving space (and hence money), patches are presented in Xpansions in a compact layout. This article describes the "mold" we use to format the patch descriptions. It's a bit cryptic at first glance, but it's pretty easy to follow. Just keep in mind that the format follows the Xpander's displays and things should make sense. Save this article for future reference, since we will not print it in every issue. When you submit patches for publication (you are working on some patches for us, aren't you?) please use this format-it saves time and effort in publishing.

To further the confusion, the patch format has been altered slightly from its old form. It's been mentioned to us by a number of people that our old format is in direct contravention to the familiar form of the manufacturer's MIDI Spec sheet, which uses "X" to indicate a feature not implemented and "O" to indicate one that is. For that reason, we have fixed our format to match this convention, so now an "O" means that a feature is enabled and an "X" means it is not.

MIDI Delay Within a Multi Patch
by Glenn Workman

These are some thoughts on using a Multi Patch to create a MIDI delay line. Create a version of the "echo" sound by setting all it's ENV delay values to the same amount. For example a setting of 43. is about 500 ms. of delay. Then select FREERUN on page 2 of those ENVS. On some sounds this may require some tweaking for the other values, but FREERUN is necessary for the ENV's to trigger if playing staccato notes. Next set up a Multi with the "original" sound layered with the "echo" sound. That's it.

1. The "original" sound and the "echo" sound can be the same timbre or completely different. For me this is the main use of this trick. An audio delay line can echo its input but not with a new sound. Brass echoed by Flute, Strings echoed by Bells, Piano echoed by...

2. Use just the "echo" sound but MIDI'd to another synth. K5 echoed by Xpander, DX7 echoed by Xpander, Prophet echoed by ...

3. Use this technique on just one VCO and build the echo into a SINGLE patch. On the Patch page, REZCHIM is an example of this.

4. Build two (or more) "echo" sounds with successively longer delay-times and layer the whole mess. Brass echoed by Flute echoed by Strings echoed by Bells echoed by Piano echoed by... [intentionally left blank]

A Programming Discovery
by Dan Barrett

I have just been shown the most mind-blowing Xpander modulation. I wish I had discovered it myself, but I can't take credit for it. It was shown to me by Craig at the Guitar Exchange in Baltimore (a co-worker of XUG member Glenn Workman), who said he learned it from someone else. I took this "trick" and figured out what is going on; it is a powerful, general modulation technique! Here's the trick I was shown. My own analysis follows.

Many people know how to get an inverted keyboard" effect on the Xpander: just turn off KEYBD on the VCO Page 2 and modulate VCO FREQ negatively several times. But I have just been shown how to do it without using any of the Xpander's 20 modulations per patch. That's right: NONE of em! Interested? Here's how:

Create a default voice OBERHEIM by holding STORE and hitting CLEAR.

TRACK1 : KEYBD .47, 39. 31 , 23 15
VCF/VCA : FREQ = 127 .
VCO 1 , 2 : (Adjust FREOs to taste),

That's it ! My eyes popped when I did this myself at home. I couldn't believe it. It's the coolest Xpander routing I've seen in several years.

Here's the patch, graphically: KEYBD->TRACKI->LAG->VCOs. Inverted keyboards are only one of the tricks you can do by putting Tracking and Lag between any modulator and the VCOs or VCF. The secret here is the use of Lag.

This'II certainly teach fellow Xpander-ites to check out those hardwired Page 2 modulations more carefully ! Personally, I'd been ignoring them for the most part. Serves me right!

This ideas are from another source:

From ucivax!jarthur!!usc!rpi!!m2c!!dime!barrett Thu Aug 29 11:35:37 PDT 1991
Article: 21161 of
Path: ucivax!jarthur!!usc!rpi!!m2c!!dime!barrett
From: (Daniel Barrett)
Subject: Re: Portamento (Was: Microwave's need for outboard effects)
Message-ID: <>
Date: 29 Aug 91 01:58:22 GMT
References: <> <> <1991Aug28.205606.1810@Times.Stanford.EDU>
Reply-To: (Daniel Barrett)
Organization: University of Massachusetts, Amherst
Lines: 37

you can make the filter frequency glide with key presses, or smooth out sharp LFO transitions.

you can use any controller. For example (this is really weird now), you can use portamento to open the output amplifier (VCA) so the sound fades in, and then control this with a pedal. Thus, you can play a chord (silently) and then press the pedal, and then the chord plays. I used this trick to turn my sustain pedal into a bass drum.

Due to the Xp/M12's many key assignment modes, the portamento can act in a number of different ways. It can glide normally from chord to chord, or jump from outer space down to the note you are holding.


| Dan Barrett -- Grad student, Department of Computer & Information Science |
| University of Massachusetts, Amherst, MA 01003 -- |