Artifacts of the Past, Modulated into the Future

With Phase-4, we wanted to partner the richness of modern technology with some early, somewhat forgotten digital synthesis methods. Luckily, ideas that yield good results have a long shelf life so working with phase is now as fresh as ever. But before we go further, let's hear a little of what is possible with Phase-4

Phase is the Place

Anyone interested in audio synthesis starts with the classic trinity of sound descriptors: pitch, timbre, and loudness. Often, these ideas are paired with their objective battle buddies: frequency, spectrum, and volume.

But lurking in the background is the deeper concept of phase, which best corresponds to time (the backdrop for all music and most art). Operating at this level allows you to uniquely alter waveshapes being produced, either while maintaining their frequency or severely manipulating it. So while phase is often misunderstood, its effects can register loud and clear to the listener.

The methods used by Phase-4 were born at a fractious moment in the history of music technology. So let’s take a moment to revisit that transitional time.

Origin Stories

At the heart of Phase-4 are two techniques: phase distortion and phase modulation. Both techniques came in the early days of commercial digital synthesis. In that era, digital techniques were often compared to analog instead of considered on their own terms — a common mistake in this field.

Going from analog to digital audio was so striking because the sounds offered by each technology were so different. Analog synthesizers relied on thick, beefy oscillators that were sculpted down by creamy filters.

By comparison, digital tools were truly thin and underpowered. But they were also highly efficient compared to analog’s power-hungry modules. Put more plainly, early digital synthesizers were true infants: newly born, of limited capacity, and awaiting instructions on how to behave. Due to their finite resources, the name of the game was to do more with less. And once techniques like phase distortion and phase modulation were invented, digital hardware started providing both new and precise sounds to complement analog’s strengths.

Logistics Make the Difference

While it is easy to get excited by the synthesis systems and filters being used, the infrastructure of any instrument determines how usable, deep, and practical it really is. So let’s take an overview of the components of Phase-4 before we get into the details.

Four identical oscillator units. Each unit is named with a letter representing its color: R(ed), B(lue), Y(ellow), and M(agenta). Each oscillator shares these features:

Flexible frequency control. Control over the pitch of each oscillator is applied in three stages.

First, an oscillator usually tracks its pitch based on incoming note messages, but this can be disabled in favor of a static frequency.
Second, a ratio is applied to that frequency so that you can set oscillators relative to each other. Both sides of the ratio can be set so that:

  • 1:1 (or “one to one”) would keep the frequency unchanged.
  • 3:1 would raise the frequency by 300% (or musically speaking, one octave and a fifth up).
  • 1:2 would lower the frequency by 50% (or down one octave).
  • 0:[anything] would effectively zero out the frequency, an interesting effect when applying modulation from other oscillators.

Finally, two offsets can be applied. A pitch offset (in semitones) can be applied to the frequency. And detuning can also be applied in Hertz, either in mono (applied equally to the left and right channels) or in stereo (with an inverse amount of detuning applied to the right channel).

A phase distortion section. Described more fully below, this portion of each oscillator allows the creation of distinct and dynamic waveshapes.

Phase modulation options between oscillators. Also described more fully below, each oscillator can be phase-modulated in various amounts by all of the four oscillators. (Yes, that includes self-modulation, or “feedback.”)


Global oscillator controls. In addition to the individual oscillator parameters is a small bank of controls that apply to all oscillators. Parameters here include PITCH offset and GLIDE time, as well as global SHAPE and MOD controls, which provide proportional control over the phase distortion and phase modulation amounts within each oscillator (more on that below).

A modified multimode filter. A seven-mode resonant filter, with keyboard tracking, drive, feedback, and a dedicated filter envelope to modulate the cutoff frequency either upward or downward. On top of all that, the cutoff frequency can also be modulated in various amounts by each of the oscillator units, adding additional levels of depth to the filter as well.

Per-voice gain for soft clipping. In addition to standard panning and volume control for the instrument’s output, additional gain can be applied to each voice that is produced. This allows for layered soft clipping to be part of the instrument’s sound before any channel audio effects are even considered. When coupled with the ample oversampling happening under the hood, each voice of Phase-4 is well off from the start.

Taken individually, each of these components can produce interesting results. Taken together, Phase-4 emerges as a instrument both rich in subtlety and capable of indulgent excess.

Let’s focus on just a few of these sections and how they turn a small amount of input into limitless possibilities.

Distorting Phase, Reshaping the Wave

Phase distortion was pioneered by Casio in their CZ-series instruments. This technique took basic waveforms (like sines) and then allowed you to bend the read angle, effectively transitioning from simple to rich waveforms in a reliable fashion. Phase-4 builds on these ideas, providing a set of individual parameters that interplay in unique ways.

The SHAPE parameter continually transitions from the current algorithm’s original waveform (when set to minimum) to its maximum variation. There are several algorithms available, each determining the source waveform and the path that it will traverse as the SHAPE setting is adjusted.

Complimenting each algorithm is a formant control. Set from 1 to 9, this value represents the harmonic being emphasized within the oscillator’s tone. (Settings above 1 define additional sine cycles to be inserted into the original waveshape.)

Finally, the phase (°) control uniquely alters the waveshape. Rather than just offsetting the signal, it also shifts the portion of the waveshape used for phase distortion. This creates additional variation in the waveforms produced by each oscillator.

Taken together, the phase distortion portion creates subtle adjustments to each oscillator’s waveshape. Whether you are setting values for each oscillator that sit well together, modulating select parameters with Bitwig Studio’s dynamic modulation system, or turning things to eleven and creating controlled chaos, the range of sounds here is profoundly usable.

The results are what you would expect from phase distortion: simple source materials transformed into dynamic waveshapes and sounds.

Phase Modulation, by Any Other Name

Taking basic sine wave oscillators tuned in musical intervals and getting them to manipulate each other (via modulation) can transform a very simple waveshape into more complex variations and spectra. And so linear frequency modulation (FM) synthesis became synonymous with Yamaha and their pioneering DX7 synthesizer. But as usual, branding can be both helpful and semi-accurate.

As is now generally acknowledged, Yamaha’s early digital synthesizers were not actually performing frequency modulation. Rather, they were using phase modulation. Since phase represents the position of a waveform, modulating phase produces an effect similar to frequency modulation but by different means.

Well, Phase-4 is not afraid to call it what it is: phase modulation. Each oscillator has four colored knobs, each setting the maximum amount of phase modulation allowed from the oscillator of that color. (The knob in the local oscillator’s own color creates a feedback effect, routing the oscillator’s output back to its own modulation input.)

These attenuators work along with the oscillator’s MOD parameter, which scales all four of the phase modulation levels proportionally. So if the big MOD knob is set at maximum, the individual modulation amounts are all applied at full value. And if MOD is set exactly to 50%, the modulation amounts applied to this oscillator will each be half of their attenuator’s value.

So multiple sources of modulation can be fine tuned, and the master MOD control moves them all together. Even before trying intermodulations (R modulates Y, who modulates R…), the MOD buss provides an expressive tension–release option for spectrum. Rather convenient that Bitwig Studio has an arsenal of modulators, ready for any musical situation.

A Last Word on the Global Controls

The SHAPE and MOD parameters on each oscillator control the respective amounts of phase distortion and phase modulation being applied to that unit. And so the SHAPE and MOD global controls act as proportional masters of for each of the four oscillators, providing yet another way to shape many layers of sound at once.

Additionally, the SHAPE and MOD global controls are mapped to the X and Y axes, respectively. With a single mouse gesture, you can click on the 4 ball icon and move both the SHAPE and MOD controls. This should satisfy both sound designers who enjoy macro controls and anyone who is just hungry for power.

Similar to how the individual phase modulation attenuators of an oscillator work with its MOD master parameter, each oscillator’s SHAPE parameter represents the value used when the global SHAPE control is set to maximum (100%). Conversely, the global SHAPE control set at minimum will allow no phase distortion (0%) on all oscillators. And this is handled similarly for the relationship between oscillator MOD parameters and the global MOD control.

If an oscillator’s SHAPE or MOD parameter is set below its maximum, an additional ball icon with the oscillator’s color will appear on the X–Y pad. Each of these balls can be clicked and moved to adjust that oscillator’s SHAPE and MOD parameters.

Whether you use these global controls or not, you can always leave those controls at maximum and then use the X–Y pad for setting the individual oscillators via their colored balls. The workflow opportunities are pretty rich.

The Sum of All Modulations

Both phase distortion and phase modulation techniques accomplish the trick of producing dynamic results from minimal source materials. Synthesizers are at their most interesting when their sound changes over time. If an instrument is capable of subtle depth, then it can likely go deep as well.

Phase-4 is informed by good ideas from the past and advanced technology from the present to support the music you are ready to make. Welcome to your sonic palette’s next phase.

Hear Phase-4 in Action

Here's a selection of demo songs, showcasing what Phase-4 can do.