Most synthesizers and samplers these days come with hundreds of preset sounds. When combined with the vast number of user-created preset banks floating around the Internet and the array of expansion cards available for many hardware models, you may wonder why anyone would bother to learn how to program one of these beasts. The answer, of course, is originality, and it's a lot simpler than you might think to tweak your way to new sounds that will set off your next masterpiece.

In this article, I'll take an operational approach to synthesizer programming by exploring the quickest route to customizing factory presets. Our starting point will be the General MIDI (GM) sound set, which contains 128 sounds covering all of the basic categories. Most synths and many samplers contain a bank conforming to the GM standard. But if that doesn't include your model, you can still follow along, because everything I cover here will apply in almost any context.

One thing you will definitely need is a programmable instrument of some sort. That can be a hardware or software synthesizer or sampler of just about any design. If you only have a preset synth (such as the Yamaha CBX-K1XG), you may still be able to get some mileage out of it if it allows MIDI or built-in controllers to alter basic preset parameters. I'll refer to that option as we go along.

GETTING UNDER THE HOOD

The first thing you need to do — which is also often the biggest hurdle to overcome — is to learn how to get into your module's patch or program editor and find the various settings you want to adjust. If it's a hardware synth of fairly recent vintage, it will probably have an LCD, an array of buttons for navigating various modes and menus, and one or more knobs for adjusting settings. If you're lucky enough to have a large LCD screen or a software editor that runs on your computer, things will be much simpler. If not, there's unfortunately no way around stepping through multiple menu layers trying to decipher cryptic parameter names like “VDA1 EG” and “AT41 AL99 DT93.”

Needless to say, your manual is your friend — it's your only way under the hood. Here are a few things to look for as you browse through your unit's documentation.

If your synth has several modes of operation, find out how to select the mode that plays a single sound on a single MIDI channel while editing. That is often called Patch or Program mode. Once in that mode, you will need to enter the program editor, for which there is typically an Edit button. In the editor, you'll need to learn how to step through menu pages to find the parameters you want to edit, how to move among the multiple parameters that occupy the same page, and how to change selected values. There are always buttons or knobs dedicated to those functions, and using them will quickly become second nature.

Most hardware models store a large number of factory programs in ROM (that can not be edited) and have a smaller user area of RAM for programs you create. You will need to dip into the manual to learn how to move, copy, and save programs in the user area. Otherwise, all your hard work will be wiped out when you turn the machine off. If you're working with a software device, you need to remember to save your work — preferably to a new location or using a new name so you don't overwrite the original program.

FIG. 1: This figure shows five common ­envelope shapes. The ADSR shape (top) is the generic synthesizer envelope shape. The Piano/Guitar envelope is characteristic of plucked and ­hammered string instruments, the sounds of which are percussive and fade out over time. The Organ envelope has no variation in level; it is either on or off. The Bow/Wind envelope takes some time to reach full level and ­sustains there until the note is released. The Bouncing Ball envelope represents a multistage ­envelope that might be used for sound effects.

PUSHING THE ENVELOPE

The quickest and easiest way to change the character of a sound is to alter its amplitude envelope. One way to think of an envelope (also known as an “envelope generator” or “contour generator”) is as a type of built-in automation that is initiated whenever a note is played. Envelopes can be used for many things, one of which is to control the amplitude (loudness) of the sound being played.

Envelopes can be described in terms of stages consisting of levels and times to reach those levels. The most common envelope — and the kind that is usually used for amplitude — has four stages, named Attack, Decay, Sustain, and Release (ADSR for short). In its simplest form, an ADSR envelope has four user settings: attack time, decay time, sustain level, and release time. The other four settings are fixed: the attack level is the maximum envelope level; the decay level is the same as the sustain level; the sustain time is the time that the note is held down; and the release level is zero. In short, when a note is played, the sound rises to its maximum level then falls to the sustain level where it stays until the note is released. The sound level then falls to zero in the release time.

A more intuitive way to picture an envelope is by its shape. Fig. 1 shows a generic ADSR envelope shape along with envelopes for several familiar sounds. Software editors and hardware devices with large LCDs usually allow you to edit envelopes graphically, making the process much simpler. But even if you are consigned to doing it numerically, it's well worth exploring the envelope settings your equipment offers.

For example, starting with a piano sound (GM preset 1) and reducing the decay time produces a damped-string effect. Increasing the release time simulates playing with the sustain pedal down. Reducing the decay time to zero and increasing the attack time significantly gives a reverse-piano effect. Increasing the sustain level to maximum and increasing the attack time a little yields a bowed-string sound. (If you're working with a preset-only synth, MIDI Control Change messages 72, 73, and 80 can often be used to control the release, attack, and decay times.) Audio Example 1 uses four sounds derived from a piano program by modifying only the amplitude envelope.

Modern synths, especially software ones, often extend the basic ADSR concept in two ways: they provide more stages and they offer control over the shape of each ramp. With the exception of the Organ, the envelopes pictured in Fig. 1 all have curved ramps, which best match the behavior of acoustic instruments.

Many synths allow MIDI Note Velocity to affect both the envelope levels and times. That allows you to play much more expressively using your MIDI keyboard. With piano sounds, for example, having the attack level (or the overall envelope level) and the decay time increase for higher Velocities gives a more realistic keyboard feel.

LESS IS MORE

In the early days of synthesis, a synthesizer's sound was characterized by filters more than anything else. With the much-expanded sound palette of today's models, the filter may be slightly less important, but it is still a key element in sound design.

Synthesizer filters are characterized by how they affect different parts of the frequency spectrum. Lowpass filters (like the treble control on your stereo) reduce the level of higher frequencies while leaving lower frequencies unchanged. Highpass filters do the opposite, and bandpass and band-reject filters reduce the level of frequencies inside or outside of a frequency band. Lowpass filters are the most common, and if your synth has only one type of filter, that is what it will be. However, it's not unusual for a synth to have a filter that is switchable among the four modes just mentioned, or even to have several filters of different types that can be arranged in series (operating successively) or parallel (operating simultaneously).

FIG. 2: The passed (blue) frequency regions of the common resonant filter types are shown here. Frequency ­increases from left to right, and level increases from bottom to top in each filter graph.

The filter setting over which you will always have control is the cutoff frequency. For lowpass and highpass filters, that is the frequency at which the signal level is reduced by half. For bandpass and band-reject filters, it is the center of the band. For lowpass and highpass filters, there is usually a resonance setting as well. That determines how much the signal is boosted (if at all) just before the cutoff frequency. Bandpass and band-reject filters sometimes have a Q setting that controls the width of the affected band. (If you're working with a preset-only synth, MIDI Controller numbers 71 and 74 can often be used to control the resonance and cutoff.) Fig. 2 illustrates the four common filter shapes and the effect of resonance.

Fixed settings for filter cutoff and resonance allow you to color the sound much as you would with tone controls or a graphic equalizer. However, things don't become interesting (and synthy) until you start changing those settings in real time. The most common tool for that job is an envelope. Usually there is an envelope dedicated to the filter with settings identical to those previously discussed for the amplitude envelope. Audio Example 2 adds resonant-filter enveloping and other modulation (more on that later) to the sounds in Example 1.

A filter's effect varies with the pitch of the sound being filtered. For that reason, you will usually find a setting called keyboard tracking or pitch tracking in the filter section of your synth. It determines how much the filter cutoff frequency is affected by the pitches being played. The keyboard-tracking range can typically be varied from zero (no tracking) to two (cutoff increases twice as fast as pitch). On some synths, it can also be inverted, causing the cutoff frequency to move down as the pitch moves up.

Keyboard tracking may not seem like a big deal, but with careful adjusting it can add life to a dead sound or it can smooth a raspy, too-bright sound. It is also useful for very-high-resonance filter effects in which you actually hear a tone at the filter's cutoff frequency. For an example, listen to GM preset 122, which produces a keyboard-tracking whistling effect using noise as a sound source.

IN THE BEGINNING

So far, we've looked at envelopes and filters for controlling the contour and frequency content of a sound. That, of course, assumes we have a sound to control. For that, your synth will have one or more sound generators, most likely referred to as oscillators, tone generators, or wave generators. The output of the sound generators might be mixed and processed by a single filter and amplifier (with a single set of envelopes), or they might each have their own signal path including filters, envelopes, and amplifiers. In the latter case, each signal path will probably claim a note from your overall note count. (For details, see the sidebar “Notes, Layers, and Channels.”)

Oscillators work in one of two ways: they generate “synthetic” waveforms, or they play samples. (Although a bit of an oversimplification, that covers most of the bases.) In either case, you can select the waveform or sample to be played. In the case of oscillators that generate waveforms, you'll have fewer initial choices, but you'll have settings with names like symmetry, pulse width, and sync that give you additional control of the sound. In the case of sample players, you'll typically have a large selection (in the hundreds) of sounds, but fewer ways to manipulate them. Let's start with waveforms.

FIG. 3: Three standard synthesizer waveforms are pictured on the left with asymmetric versions shown on the right. In the case of the square wave, the asymmetry setting is called pulse width, and the waveform is alternately called a pulse wave.

THE SHAPE OF THINGS

Fig. 3 shows three standard oscillator waveforms on the left (sine, triangle, and square) and the result of modifying their symmetry (sine and triangle) or pulse width (square) on the right. Changing a waveform's symmetry or pulse width alters its harmonic content, and generally results in richer textures.

Oscillator sync, or hard sync, is another commonly available waveshape-altering process. In that process, one oscillator, called the slave, is forced to restart its waveform in sync with another oscillator, called the master. With hard sync, the master oscillator controls the pitch. Changes in the slave oscillator's pitch setting don't change its pitch; rather, they cause its waveform to be truncated at different positions, thus affecting its tone. The mix of the master and slave oscillators in the audio output determines how pronounced the effect is.

You can use the symmetry, pulse-width, and hard-sync settings on your synth to greatly increase the variety of sound sources, but applying those effects dynamically is even more interesting. For that purpose, you'll typically find envelopes dedicated to the oscillators, or at least a way to route the filter envelope to the oscillator. You'll also almost certainly have a low-frequency oscillator (LFO) available for modifying all those settings. (If you're using hard sync, apply the envelope or LFO to the slave oscillator's pitch.)

An LFO is an oscillator that operates at frequencies that are below the audio range — typically from 0 to 20 Hz — and can be routed to change various settings. (Using one process to control the settings of another is called modulation.) Because it is operating at a low frequency, the changes it produces are directly perceivable. If you modulate the same settings with an audio-rate oscillator, you will perceive a change in the harmonic content of your sound.

LFOs generally have fewer waveform choices than audio oscillators, but you'll always find a sine or triangle shape for vibrato and tremolo-like effects. Generally, there will also be a pulse shape for gating effects, ramp-up and ramp-down (sawtooth) shapes for repeating attack and decay effects, and a random form for sample-and-hold effects. Audio Example 3 is a fat chord with LFOs and envelopes applied to symmetry, hard-sync slave pitch, and filter cutoff. LFOs can often also be used to retrigger envelopes. Look for that feature on your synth — it will open up a world of step-sequencing possibilities.

If your synth's sound generators play samples, you will not find the same shaping controls that are available for waveform generators. However, you will probably still find settings for pitch modulation by an LFO and envelope, and you will have a larger variety of sounds to start with. Both sample and synthesis-based devices also typically offer one or more effects processors for modifying sounds. See “Square One: Multi-Effects 101” and “Square One: Multi-Effects 102” in the June and July 2000 issues of EM for a detailed introduction to effects processors.

STEPPING UP

You've now had at least a brief look at most of the sound-programming features you're likely to run into on the majority of commercial hardware and software synthesizers. Getting into and finding your way around the synthesizer's program editor is the hardest part. Once you've taken the time to do that, a nearly unlimited sound palette opens up to you, and the best part is that they're your sounds.

Beyond programming new sounds, a familiarity with the inner workings of your synth provides a broad range of performance possibilities. Instruments with keyboards and onboard controllers (wheels, joysticks, ribbons, and so forth) usually allow you to route those controls to many of the settings provided in the editor. Most synths let you route external MIDI Control Change messages similarly. Whether you use your instrument live or with a MIDI sequencer, you can increase its expressiveness by taking dynamic control. So find the hood latch, get busy, and don't forget to save your work.

Len Sasso can be contacted through his Web site at www.swiftkick.com.

NOTES, LAYERS AND CHANNELS

Most hardware manufacturers proudly advertise their unit's polyphony, which is the number of notes the device is capable of playing simultaneously. Though that is a valuable measure of a unit's capabilities, and more is definitely better, it is important to be aware that playing just a single note can potentially absorb multiple notes of your synth's polyphony. That can happen in several ways.

On many models, a single patch can use two or more sound generators in such a way that each individual sound source accounts for one note of the overall polyphony. For example, if you play a patch that consists of four sound generators on a device that offers 64 notes of polyphony, you'll only be able to play 16 simultaneous notes for that program. To improve things, many models have smart “voice-management” routines that, for example, release a note when its amplitude envelope has fallen to zero, even though the note is being held.

Most models also have modes designed to let you layer programs and to play different programs on different MIDI channels. Those usually go by names like Combi, Multi, and Performance. Needless to say, those modes quickly consume the available notes. If you layer two programs in the above example, you're down to eight available notes of polyphony, and many synths can layer four or more programs.

Finally, if you're using a sequencer with your synth and sequencing parts on several MIDI channels, keep in mind that the number of available notes is split among the different channels. Fortunately, all synths are smart enough to dynamically allocate the polyphony, so you don't have to decide in advance how many notes to assign each channel. However, it's important to keep the polyphony in mind. Otherwise, adding a lead track on a new MIDI channel can result in notes unexpectedly dropping out of your lush string pad.

Beyond carefully tracking the note count in your sequences and each of your programs, many synths offer options for controlling what happens when you reach the polyphony limit. Some models allow you to reserve a minimum number of notes for a particular MIDI channel. Another option is voice priority (often called note-stealing priority or dynamic voice allocation). That allows you to specify what happens when playing a new note exceeds the limit. Typical choices are last-note priority (earliest played note is turned off), high- and low-note priority (lowest or highest note is turned off, respectively), and loudest-note priority (lowest-level note is turned off).