All too often, the sounds produced by synthesizers are lifeless and boring. (I'm a huge fan of synthesizers, so I'm allowed to say that.) Various techniques are used to bring the sound of a synth to life, and one of the most important of them is modulation.
FIG. 1: e-Lab Obsession''s modulation matrix has radio buttons for choosing sources and destinations. The LFO module occupies the same panel space but isn''t part of the matrix.
With modulation, a control signal is used to change the behavior of one or more of the synth's components in a particular way. The control signal might, for instance, raise the cutoff frequency of a lowpass filter, which would cause the sound passing through the filter to become brighter.
Most synthesizers have several modulation sources that produce control signals and have a number of destinations to which modulation can be applied. Typical sources include envelope generators, low-frequency oscillators (LFOs), and MIDI Control Change messages (see Web Clip 1). The list of destinations varies from one instrument to another and will likely include, in addition to filter cutoff and the pitch of individual oscillators, oscillator pulse width, filter resonance, effect wet/dry mix, LFO amount and speed, the mix of oscillator signals being sent to the filter, and perhaps the rate of individual envelope segments (attack, decay, and so on).
Setting Up Modulation
For modulation to occur, you need to set up a modulation routing, which is composed of a control-signal source, a destination sound parameter, and the amount of modulation. You can think of a modulation routing as being like a hose: to use it, you first need to attach one end to a water source and the other end to a water destination, and then you need to turn the tap to regulate the amount of water flowing through the hose. The manufacturer of a synth (hardware or software) determines which front-panel controls are used for setting up modulation routings, and although those controls differ from one synth to another, the underlying concept is the same.
Broadly speaking, manufacturers use two methods for setting up modulation routings, and each has its advantages. On some instruments, the routings are hardwired (the term is left over from the days when synths were constructed with actual wires rather than with computer code). An LFO, for instance, may have its own set of switches that let you choose between the possible modulation destinations. A filter may have a fixed set of modulation inputs (from an envelope, an LFO, the keyboard, and perhaps Velocity). The advantage of hardwired modulation routings is that they're easy to understand. But if none of the included routings does what you need, you can't “rewire” the synth.
Other instruments have a set of 8, 16, or more general-purpose, programmable routings, and it's up to you to determine what the routings are used for. This type of setup, called matrix modulation, is inherently more flexible than hardwired modulation. It lends itself to subtler and more visionary forms of sound design. But it's also harder to work with. When a synth has multiple panel pages, for instance, the envelope generator may be on one page, the filter on another, and the modulation routing that connects the two on yet another. That is a recipe for extra button pressing or mouse clicking as well as for added confusion. If the sound you're hearing is behaving in an unexpected way, you may have to dig a little to figure out what's going on.
Some synths have both a modulation matrix and several hardwired modulation routings. The latter are used for the most essential forms of modulation, such as MIDI CC 1 (Mod Wheel) to LFO amount.
The Matrix: Revealed
A typical modulation matrix (see Fig. 1) has fields in which you choose the source of the modulation signal (the input) and the destination (the output). Between the two is an amount slider. Using the matrix is easy: just choose a source and a destination, and then adjust the amount slider while playing the keyboard to hear the effect of the modulation. But though the concept is simple, it's worth looking at several subtleties.
First, modulation signals are potentially bipolar. That is, the value of the signal can be either greater or less than zero. A signal that drops below zero may seem absurd: how can you have less than no modulation? In fact, the idea makes sense. If, for example, an envelope generator's output is applied to oscillator pitch, you may want the pitch to rise as the envelope rises, or you may want the pitch to fall as the envelope rises. For that reason, the amount sliders in a good modulation matrix will have zero in the middle and let you set either a positive or a negative modulation amount. When the amount is less than zero, the modulation signal is inverted.
FIG. 2: The modulation matrix in Camel Audio Cameleon 5000 has minimum- and maximum-amount sliders. Here, the High setting is lower than the Low setting for the fourth routing, so the modulation will be inverted.
In some synths, however, the modulation signal is inverted in the source module (usually an LFO or envelope generator) before being sent to the matrix. That design is less flexible, because if you invert the signal being sent from that source to one destination, you'll wind up inverting it for all destinations. The source module may also have its own output-amount parameter, which means the modulation amount can be programmed in two different places. That opens the door to confusion.
Instead of having just one amount slider, some mod matrices have minimum- and maximum-amount settings (see Fig. 2). The minimum setting will be sent when the modulation signal is at zero, and the maximum setting will be sent when the modulation signal is at its greatest value (usually 127). The amount of modulation is reduced (scaled) by setting the minimum and maximum values closer together. When the minimum is set to a value that's greater than the maximum, the modulation signal is inverted: a lower input produces a higher output, and vice versa.
I've already mentioned one of the most common forms of modulation: MIDI CC 1 to LFO amount. That is used for controlling vibrato depth from a mod wheel. But how do you set that up in a modulation matrix? Unless the LFO has its own amount parameter, which can be chosen as a destination, it would seem you couldn't do it. Many modulation matrices, however, provide an easy work-around: the amount parameter of one modulation routing can be chosen as a destination for a different routing.
In that case, you'd set up one routing with your vibrato LFO as the source and oscillator pitch as the destination. The second routing would have CC 1 as the source and routing 1 as the destination.
Secondary modulation opens up a lot of musical possibilities (see Web Clip 2). For instance, you can control LFO amount from an envelope generator, which is useful for vibrato that deepens gradually after the beginning of the note. By routing the LFO to panning and using Velocity as the secondary modulation source, you can increase the depth of the panning by striking the keys harder. That might not seem very exciting by itself, but when used in combination with other modulations (such as Velocity to LFO rate), it can produce useful musical effects.
Trying out random connections in a modulation matrix will sometimes produce happy surprises. But ultimately, a better way to use modulation is to start with an idea. Figure out what changes you want to make in your sound, and then choose the ingredients that will do the job. Modulation can give tired sounds a new lease on life.
Jim Aikin writes about and has fun with music technology in his home studio in Northern California. He also plays electric cello in a jazz/folk quartet called Night Harvest with ace violinist Jim Hurley (of Dan Hicks and of Queen Ida).