Proteus Power

An insider's tips on maximizing this monster module.In recent years, sample-playback modules have acquired a reputation for being relatively boring devices,

An insider's tips on maximizing this monster module.

In recent years, sample-playback modules have acquired a reputation for being relatively boring devices, capable of producing little beyond the usual meat-and-potatoes piano, string, and brass sounds. So it hardly seems fair to categorize E-mu's Proteus 2000 as a sample-playback unit. True, its sounds are based on samples stored in ROM, but that's where its similarity to most playback appliances ends.

The synthesis capabilities of the Proteus 2000 easily overshadow those of lesser devices and offer the synthesis aficionado a huge bag of tricks for seriously warping sounds. I'm going to touch on a few of my favorite tips and tricks, but this article is by no means an exhaustive survey of what this powerful machine can do - it's really just a start.

CONSTRUCTION KITApart from the Instrument kit GM Dance, the Proteus 2000 does not provide a standard General MIDI drum kit. This might force you to edit some GM sequences if you wish to play them with anything other than the decidedly electronic textures of the GM Dance Instrument. You can, however, do a pretty good job of creating a variety of GM-compatible kits with the Proteus 2000's extensive layering and linking capabilities. By using all four Layers in each of three linked programs, you can create a preset with 12 separate Layers, each of which can contain a chunk of your customized GM drum kit.

Here's an example: Select preset 079 (Acoustic 2) in CMPSR bank 3. As you play up the keyboard, starting on MIDI note 35 (B0), you'll discover that the kit corresponds to the GM layout for only four notes. I'm not going to worry about assigning the notes below note 35; they aren't used very often. (If you need them, you can extrapolate from the rest of this exercise to determine how to assign them the way you want.)

When you hit note 39 (D superscript #1), you'll find a snare sound where handclaps would be in a GM-compatible kit. To correct this, simply add a new Layer beginning at note 39. In Edit mode, set Layer 1 (which is providing the first group of percussion samples) to sound only from note B0 to D1:

Now select Layer 2, and assign it to sound only on D superscript #1:

While Layer 2 is still selected, scroll to the Instrument page and change the assignment to one of the handclap sounds, such as Claps 2:

This Layer has been transposed down 24 semitones, so you'll probably want to change that:

Because the claps are now in a separate Layer, you have the freedom to tune, pan, mix, and process them separately from the other sounds in this kit - things you normally wouldn't be able to do with a drum kit that uses a preset multisample. For example, using Send2 for the claps allows you to splash more reverb on them than you would use for the rest of the kit. You can assign Send2 to Layer 2 on the Mix Output page:

After adding more Layers to this preset, you can adjust the reverb amounts for each Layer (up to four Layers per preset) from the FXA Send Amounts page:

To complete the GM drum kit, continue assigning Instruments to the appropriately split Layers. For example, you can assign Instrument 0630 to Layer 3:

This Instrument corresponds to the GM standard all the way to note 51 (D superscript #2), so we'll split it like this:

For the fourth Layer, we can use the GM Dance Instrument, which provides relatively generic-sounding percussion instruments in a GM-compatible map up to A superscript #3:

This example illustrates the basic approach, but you can certainly take the idea much further. If you need additional Layers, simply link another preset (or two) to this one. Each additional preset provides four more Layers to mess around with. You might want to use the additional Layers to cover areas of the keyboard above and below the ranges I've dealt with here. Also, having specific percussion sounds assigned to individual Layers can provide some flexibility in tailoring the drum kit to a specific situation. For example, if you assign a cymbal to its own Layer, you can adjust its decay time, reverb amount, and panning without affecting other sounds in the kit.

TUNE-UP MASTERI'm also fond of employing the user tuning tables to control the pitch of individual drums in a Layer. This lets you tune the toms, for example, to get additional pitches that aren't preset in the stock drum map. To try this technique, set the Instrument's keyboard tuning to the User 1 tuning table:

Then hit the Master button and scroll to the tuning-table definition page:

Underline the key number and play the low tom (F1) on the keyboard. Underline the Crs value and set it to 038. This tunes the low tom several semitones lower.

You can repeat this procedure for any drum sound you'd like to tune.

BENDY-BOYThe idiomatic "dive-bomb" sound associated with electric guitar can be difficult to reproduce on most synths. That's because most synths let you set only a single value for the Pitch Bend range. In other words, if you want to set the bender to drop a fifth, you end up with the Pitch Bend range going up a fifth as well. That can make it hard to perform the guitar's characteristic half- and whole-step upward bends. Fortunately, you can set the bender in the Proteus 2000 so it bends down by a different amount than it bends up.

The trick is to use a Proteus modulation processor called the Diode, which passes only positive input values. By routing the Pitch Wheel through the Diode and then to the Pitch parameter, you can add a greater modulation amount to the Pitch Bend range, but in one direction only.

The following patch requires two PatchCords (modulation connections). From the Edit menu, route one PatchCord from the Pitch Wheel to the Diode.

Then connect the Diode to the Pitch parameter.

Set the modulation amount according to the effect you want. For example, if you set it to -010, and the Pitchbend Range for the patch is set at ñ2, the bender will bend notes up a second and down a fourth. If you set the modulation amount to -016, the bend down will be a fifth. To use the Diode output to control the upward bend range, set the first PatchCord to +100:

WAVE OPSTo paraphrase a well-known maxim, you can never be too rich or too thin, or have too many waveforms. Even though the Proteus 2000 has nearly 1,200 different waves, there may still be times when you want something that isn't in the instrument's ROM. Luckily, there are a couple of easy ways to produce new instrument sounds for your custom presets.

One of the easiest methods is to experiment with a sound's start time. Setting the start-time number to something greater than 0 starts the playback of the sample at a later point in the waveform. This technique may not make much of a difference when applied to a synth-based waveform, but if you apply it to an acoustic-instrument sample, you can lop off the attack of the sound and leave just the sustaining portion. Because we identify an acoustic sound largely by its attack portion, removing the attack can leave you with a waveform that retains some acoustical characteristics but isn't readily identifiable.

For example, select the str:Solo Quartet patch (089) in CMPSR bank 7. Hit the Edit button, scroll to the Sound Start menu, and set Start to 127:

This setting chops the gritty bowed attack off the front of the sound. Now, to further mangle the waveform, I'll apply the second trick: sample stretching.

The Proteus 2000 provides two ways to change a sample's pitch: transposing and tuning. Transposing the sample works by shifting the keyboard position of the waveform, which preserves the key range assigned for each waveform in a multisampled Instrument. Tuning shifts the pitch of the multisamples, pushing them outside their predetermined key range. By combining the two approaches, you can keep a preset in pitch while forcing its individual samples outside of their range.

For example, set the transposition for Layer 1 up an octave:

And set its tuning down an octave:

Try increasing the release time a bit and dialing up an appropriate filter - perhaps one of the Vow (vowel) filters - and you'll have a fairly sick-sounding vocal group.

UPS AND DOWNSAnother technique is triggering a Layer on key-up. This lets you create a mandolin-style tremolo with a guitar sound.

Start with a straight-ahead guitar sound, such as gtr:Steel (048) in CMPSR bank 7. Use the transposition trick that I just explained to make the steel-string guitar sound a bit more like a mandolin. Transpose the Instrument up 7 semitones, and then tune it down 7 semitones to get that slightly thunky mandolin texture. Now set up the gate to control sample retriggering:

(Note that any number greater than 0 will suffice for the PatchCord amount.)

This setup causes the Layer to retrigger when the key is released. The effect will be pretty subtle, because the envelope cuts off the sound of the Layer when the key is released, but this is easily remedied. Scroll to the Volume envelope page and make sure the Mode is set to time-based (or tempo-based, if you want to time the envelope to a particular tempo). Set the values for Attack1, Attack2, and Decay1 to a rate of 0 and a level of 100 percent. Set Decay2 to a rate of approximately 76 and a level of 0 percent. Set Release1 to a rate of 0 and a level of 100 percent (so the envelope jumps to full volume when the key is released), and set Release2 to a rate of approximately 54 and a level of 0 percent (so the note ultimately decays to silence).

If you want to get fancy, use the PatchCords to scale the decay and release times to keyboard position, so that higher notes produce quicker decays and releases, as with a real mandolin:

To complete the mandolin effect, turn on the Chorus parameter. Choose a setting that's not too heavy - perhaps around 12 percent with the width left at 0 percent.

SAXX GRUZZA saxophone growl can be a handy sound to have in your collection. Although there is more than one approach to creating this kind of sound, an easy way to get a reasonably good result is to use the Auxiliary Envelope to create the characteristic "buzz." You can then use a controller, such as Aftertouch or the mod wheel, to bring in the effect.

To create this sound, first locate wnd:Grouch (042) in CMPSR bank 0. This provides an aggressive saxophone sound as a starting point. For the basic growl, you can set the Auxiliary Envelope to repeat, and set each segment to travel a small distance at a fast rate. Begin with all the attack and decay segments set to a rate of 1, and alternate between a level of +1 percent and -1 percent:

Note that I've used positive and negative levels in equal increments. That way, the pitch modulates around 0 and doesn't get pushed sharp or flat as the depth of the effect increases. Also, the release times won't matter unless the sound you're working with continues to decay after key-up. If it does, be sure to set the level for Release1 and Release2 to +0 percent, or your note may drift off pitch when released.

Once the Auxiliary Envelope has been set, use the PatchCords to route it to a likely destination. I've found that routing it to the Pitch parameter works well, but routing it to the filter or amplifier can also produce useful results. Whatever the destination, though, you'll probably want to hook up a controller to fade the effect in and out. To accomplish this, use another PatchCord to route the controller (I'll use the mod wheel for this example) to control another PatchCord amount. Let's hook up the Aux Envelope to control Pitch, but with a depth of 0:

Then hook up another PatchCord routing mod wheel to the Amount control:

Finally, if you want to get tricky, use at least two Layers for this patch, with a crossfade controlled by the Mod wheel, breath controller, or some other controller. This allows you to set up fades between mellow, smooth sax sounds and the grittier, buzzy sound you just created.

RANDOM RAMBLINGSWithout going into too much detail, I've set down a few ideas that might get you started in some new directions:

1. Route one of the squared-up LFO shapes - one of the Pat: shapes, for example - to control SRetrig on a percussion Instrument. You will have hours of fun as you experiment with new drum grooves based on the LFO rate.

2. Use the Sub outs for the dry sounds, and the Main outs for the wet (effects) signal. This lets you control the overall effects mix from a couple of faders on your mixer - handy for live performances.

3. Several of the Proteus 2000 filters can be made to distort, but the BassOMatic filter produces a warm, tubelike distortion. Adjust the Q parameter to control the amount of distortion, or use Velocity or another controller to crossfade between distorted and clean Layers. This is a nice effect for electric pianos and organs, among other things.

4. For sweeping pads, try using one of the vocal filters on a Layer, modulated by a slow LFO for some extra animation.

5. You can achieve some of the most interesting effects by modulating modulators. In other words, try modulating an LFO's rate and/or depth from another LFO, which might itself be modulated by an envelope or even the first LFO. Some very interesting cascading patterns can be created by routing the results of these "monster mods" to several destinations, such as Pitch, Filter, Pan, and so on. You can keep everything in time by basing LFO rates and envelope times on bpms, or you can let things run free to produce sounds that constantly change and shift.

6. Finally, don't forget to read your Proteus 2000 owner's manual, especially the programming tutorial. It's an excellent source of small exercises that can turn into big ideas. Also check out back issues of EM; the Proteus 2000 was reviewed in the September 1999 issue.

A number of helpful Proteus 2000 links are online, including E-mu's own Web site ( An active Proteus 2000 user group is at, and Jeff Donnici runs an excellent Web site at, where you can find tons of information and links.

I've often wished that I could have my own custom sound bank available in ROM, providing instant access to the sounds I use the most. With the Proteus 2000's support for sounds burned into nonvolatile Flash ROM, that wish is now a reality. That's great news for gigging musicians who want to use custom samples in their act but prefer the convenience of a sample-playback module. You can also do things with the Proteus 2000 that you can't with other modules, such as apply the amazing z-plane filters to your custom drum loops, or create your own custom drum kits with samples, mapping, panning, and tuning designed to your own specs.

The process of creating custom ROMs is relatively simple (although you'll have to get your hands on an Ultra-series EOS sampler if you don't already have one). Simply pop a Flash-ROM SIMM into the EOS Ultra, put together the bank of sounds that you want to author, save the bank to the Flash-ROM SIMM, transfer the SIMM from the EOS Ultra to the Proteus 2000, create presets based on the new ROM, and save the presets (as a bank) to the Flash ROM.

The procedure will go smoothly if you take a little time to plan your project before starting. The main thing to keep in mind is that you have a limited amount of memory. Flash-ROM SIMMs currently come in only one size: 16 MB, although 32 MB SIMMs may be available in the future. The Proteus 2000 has a total of four slots that can accept Flash SIMMs, but one slot is already used for the Proteus 2000's factory ROM banks. (You can, however, remove the factory ROM and fill all four slots with your own ROMs.)

With 16 MB SIMMs in three slots, you have a maximum of 48 MB to play around with. Remember, however, that an Instrument must be contained in a single bank. In other words, if you want to store a 20 MB multisampled piano patch in Flash ROM, you must do something a bit tricky, such as creating the lower half of the piano in one 16 MB bank and the upper half in another 16 MB bank. (you can always fill any leftover Flash memory with other sounds.) Then, when you've installed your newly authored Flash SIMMs in the Proteus 2000, you can re-create the original piano patch by layering or linking the instruments from the two different ROM banks.

Cramming a wealth of sounds into a relatively small space is truly an art, and attempting to put 16 MB to their most efficient use makes you admire the E-mu sound designers who stuffed hundreds of samples into the 32 MB of sample ROM that ships with the Proteus 2000. None-theless, anyone familiar with sampling shouldn't have too much trouble understanding the basics of efficient memory usage: never use two samples when one will suffice, do not use higher sampling rates than necessary, and keep the samples as short as possible.

The Proteus 2000's synthesis resources are quite extensive, so you might not need to sample a completed sound. You may be able to get away with sampling a component of the sound and processing it later in the Proteus 2000.

Consider, for example, the re-creation of some of your favorite synthesizer pad and sweep sounds. If the sound you want to capture is characterized by a long filter sweep, try opening the filters on the synth all the way and sampling the resulting unfiltered sound. Then see if you can re-create the sound using the extensive filtering available in the Proteus 2000. You might find that once you've turned off the filters, you're left with a relatively straightforward layer of synth waveforms - some of which may already be available in Proteus 2000 ROM. You might even be able to re-create the sound you want without making any custom samples at all, which is a great way to get to a deeper level - not just with the Proteus 2000, but with your other synths, too.

Finally, as you're putting together your banks in the EOS Ultra, keep in mind that nothing is forever. You may wish to revisit those banks to make changes at some later date, so it's a good idea to back up your original EOS banks. That way you can easily retrieve them if you need to.