Auditory Streaming Effects in Synthesizer Patch Design

While gestalt psychology is well-known in its applications in the visual design realm — since graphic and UX designers frequently apply its various principles such as Similarity, Continuation, Closure, Proximity etc. in their output — the auditory application of Gestalt is not frequently commented on.

The field of study that considers our apprehension of acoustic perceptual wholes is called auditory scene analysis or ASA. The classic text of ASA is Albert Bregman’s Auditory Scene Analysis: The Perceptual Organization of Sound. While gestalt studies visual objects, ASA studies streams of sonic events. One of the phenomena at the center of experiments in ASA are the auditory streaming effects of fusion and segregation, which has to do with whether we think of sonic components as all belonging together as part of the same stream, or whether we perceive a stream as composed of multiple sounds or streams.

For example, in this short Bach excerpt, even though there is only one line of melody, because one or a few of the notes are of a much higher pitch level relative to the others, we start to perceive the highest notes played as their own separate melodic line.

Auditory segregation in a single line Bach melody.

The characteristics required for us to determine whether we perceive a single stream or multiple streams of the same sound are studied by using ABA patterns. An ABA pattern simply repeats two similar pitches, the A and B pitch. These pitches are then experimentally varied, for instance by differentiating them in loudness, or by pitch distance etc. As shown below, a basic ABA pattern can be perceived in 6 different ways: as part of the same stream (the first row shown), or in 5 possible variations of a two perceivable streams.

Image Source

The relevance of this to the design of synthesizer patches (or presets, or timbres — the term you use may differ based on your audio application or personal nomenclature) is that often with just a light touch of sound design, it is possible to make a single simple synth voice sound like several. Listen to this example and think about how many separate synth voices you perceive.

Depending on how I focus my attention, I can hear up to four separate possible streams: the sustained notes, a separate bending note, and high/low pulses at different frequency levels. However, everyone might perceive this sound differently as there can be a lot of subjective variation in how auditory fusion or segregation takes place when we’re listening.

This perceived differentiation of a single synth voice is achieved through a very limited set of parameters. The synth used in the example above is from SimSynth, a stock plugin with FL Studio. Below I show what this customized preset looks like on its front and back end. There is a single square wave oscillator with two envelopes, one for volume and one for the filter. These envelopes are different as per usual with sound design. A square wave LFO is modulating the cutoff frequency of the filter and its rate is set to produce interesting phasing effects with the main pitch, since they are both square waves. There’s also a basic chorus activated and a retrigger. As you can see from the plugins ‘back side’ there’s no arpeggiator applied even though it might sound like some kind of arpeggiation is happening with the descending pitches.

SimSynth plugin front end.
SimSynth plugin back end.

The pattern is made up of just four long notes that repeat once, but with the occasional pitch bend used to join every other note. This alternating pitch bend can give the sense of another auditory stream in its own right, but everyone’s perceptions may vary on this. In FL Studio, if you’re using an Image Line generator plugin you can write special pitch bend notes in the piano roll.

Repeating four note pattern.
The very short notes are actually just pitch bend notations in the piano roll.

I typically find that when taking a more digital craft approach to synth voice creation (as opposed to just selecting a ready-made preset), I can usually get a sound to a very rich place with just a few minor adjustments. Often, after the first few sound design moves, the addition of more parameter adjustments tend to degrade the aesthetic quality of a sound.

A minimalist approach to synth patch design in which ASA principles are applied can allow you to create a sense of multiple synths present when in fact there’s only one. Practically there’s the benefit of less CPU load, since one synth preset of your own making is creating the illusion that there are multiple synths in the mix. In terms of arrangement, a lot of compositional work can be accomplished with just a few patches of your own design which have been shaped to be perceived as multiple auditory streams by your listener.

You can hear the particular timbre discussed above in the music track below: