Effects as Synthesis Tools

Feedback as a Sound Source

Feedback is what happens when a signal is routed back into itself. Every effect that has a feedback control — delay, reverb, flanger, phaser — is capable of self-oscillation if you push the feedback high enough. At that point, the effect is generating sound on its own. It no longer needs an input.

In a delay pedal or plugin, feedback controls how much of the output gets fed back to the input. At low values, you hear echoes that decay. At moderate values, the echoes sustain. At 100% feedback, the echoes never decay — whatever enters the delay stays there forever, looping. Push feedback past 100% (many analog-modeled delays allow this) and the signal grows with each repetition, building to distortion and then to a howling, self-sustaining tone.

That self-oscillating delay is a synthesizer. It has a pitch (determined by the delay time). It has a waveform (shaped by the character of the delay’s internal processing). It has amplitude (controlled by the feedback amount). Modulate the delay time and you get pitch bends. Modulate the feedback and you get dynamic swells.

Try creating ringing notes from delay lines by using short delays with long feedback values. The delay time sets the pitch. The feedback sustains it. A gate or VCA controls when the ringing starts and stops. Patch an envelope into the feedback amount and you get a note that rings up, sustains, and then decays as the feedback drops below unity.

Delay Time as Pitch

This connects directly to what you learned in Chapter 12 about Karplus-Strong synthesis. A delay line with feedback is a resonator. The delay time determines the resonant frequency.

The math is straightforward: frequency = 1 / delay time. A 1-millisecond delay resonates at 1000 Hz. A 2-millisecond delay resonates at 500 Hz. A 5-millisecond delay resonates at 200 Hz. As the delay time gets shorter, the pitch gets higher. As it gets longer, the pitch drops.

At delay times above about 30-50 milliseconds, you stop hearing pitch and start hearing rhythm — the individual echoes become distinct events rather than a fused tone. Below about 1 millisecond, the pitch is so high it moves into the upper reaches of hearing or beyond. The sweet spot for delay-as-pitch is roughly 1 to 20 milliseconds, corresponding to pitches from 1000 Hz down to 50 Hz.

When you mix a dry signal with a very short delayed copy, you get a comb filter. Some frequencies add up (constructive interference) and some cancel out (destructive interference), creating a series of peaks and notches across the spectrum. The pattern repeats at intervals related to the delay time. Sweep the delay time and the peaks and notches move — that is flanging.

Comb Filtering and Flanging

A flanger is a comb filter with a slowly modulating delay time. The delay is very short — typically 1 to 10 milliseconds — and an LFO sweeps it back and forth. As the delay time changes, the comb filter pattern shifts, producing the characteristic swooshing, jet-engine sound.

But a flanger with high feedback is more than an effect. The resonant peaks of the comb filter become so pronounced that they dominate the signal. At maximum feedback, those peaks become pitched — the comb filter is now a bank of resonators, all harmonically related, ringing at frequencies determined by the delay time. Feed it noise and you get a pitched tone emerging from the noise. Feed it a drum loop and the drums take on a distinct pitch coloration.

This is the same principle as Karplus-Strong, viewed from a different angle. Karplus-Strong uses a delay with feedback and filtering to model a string. A flanger uses a delay with feedback and modulation to create swept resonances. Same mechanism, different application.

In VCV Rack, you can build a flanger from first principles: take a signal, split it, delay one copy by 1-5 ms, modulate the delay time with a slow LFO, and mix the two copies back together. Add feedback from the delayed copy back into the delay input. As you increase the feedback, listen for the transition from subtle coloration to pitched resonance to self-oscillation.

Ring Modulation and Amplitude Modulation

Ring modulation creates frequencies that were not present in either input. It multiplies two signals together, and the mathematical result of multiplying two sine waves is two new frequencies: the sum and the difference of the originals. The original frequencies disappear.

Feed a 440 Hz tone and a 100 Hz tone into a ring modulator. Out comes 540 Hz and 340 Hz. The 440 and the 100 are gone. If the two frequencies are harmonically related (say, 440 Hz and 220 Hz), the sum and difference (660 Hz and 220 Hz) are also harmonically related, and the result sounds musical. If the two frequencies are not harmonically related (say, 440 Hz and 137 Hz), the sum and difference (577 Hz and 303 Hz) are inharmonic, and the result sounds metallic, clangy, bell-like.

Ring modulation works as both a synthesis technique and a vocal processing tool. Ring-modulating a voice creates the classic “robot voice” or “Dalek” effect — the harmonic structure of the voice is transformed into something alien because the sum and difference frequencies do not align with the original harmonics.

In VCV Rack, a ring modulator is just a VCA with bipolar CV input, or a dedicated multiplier module. Route two oscillators into a multiplier and listen to what comes out. Start with both oscillators at simple ratios (octaves, fifths) and then detune one gradually. Listen for the transition from harmonic to inharmonic to chaotic.

A compelling approach: combine ring modulation with de-aggregated polymetric sequencing. Give each oscillator feeding the ring modulator its own independent sequence, so the sum and difference frequencies create constantly shifting melodic patterns that neither sequence alone contains. Ring modulation as a compositional tool, not just a timbral one.

Vocoders and Cross-Synthesis

A vocoder splits the frequency spectrum into bands — typically 8 to 32 — and uses an envelope follower on each band to track how loud the modulator signal is in that frequency range at any given moment. Those envelope values then control the gain of corresponding bands on the carrier signal.

The carrier provides the pitch and raw timbre. The modulator provides the spectral shape — the vowel sounds, the consonant articulations, the dynamic contour. When the carrier is a rich synthesizer pad and the modulator is a voice, the synth appears to talk.

Vocoders connect to the broader concept of spectral shaping explored in the physical modeling chapter’s formant synthesis section. A vocoder does not transfer pitch. It transfers spectral shape — which frequencies are loud and which are quiet at each moment. The carrier needs to be spectrally rich (a sawtooth wave or noise works well) so the vocoder has content in every frequency band to work with. A sine wave carrier sounds thin through a vocoder because there is no harmonic content for the upper bands to shape.

Cross-synthesis is the more general concept. Any process that takes the spectral characteristics of one sound and applies them to another is cross-synthesis. Vocoders are the most common implementation, but convolution (multiplying spectra in the frequency domain) can achieve similar results with different character.

Distortion and Waveshaping as Timbral Tools

Distortion adds harmonics. That is its fundamental nature, whether the distortion comes from an overdriven tube amplifier, a clipping circuit, a bitcrusher, or a mathematical waveshaping function. The specific harmonics added — and their relative levels — depend on the type of distortion.

Soft clipping (tube-style saturation) adds predominantly odd harmonics with a gradual rolloff. Hard clipping (transistor-style) adds odd harmonics more aggressively. Wavefolding — sometimes described as “gaining harmonic content with unique forms of distortion” — folds the peaks of a waveform back on themselves, adding both odd and even harmonics in complex ratios that change with the input level.

In synthesis, waveshaping is a deliberate timbral tool, not an accident. Feed a sine wave into a waveshaper and you get a complex waveform. The character of that waveform depends on the transfer function — the mathematical curve that maps input values to output values. A gentle S-curve adds subtle warmth. A staircase function adds aggressive digital harmonics. A folding function creates the rich, evolving timbres associated with West Coast synthesis.

A productive exercise: pick a bass sound from any record with prominent distorted bass and identify the types of distortion used. The exercise connects distortion back to synthesis: understanding what a distortion circuit does to a waveform is the same skill as understanding what any other waveshaping process does. It is all harmonic manipulation. For more on effects processing in a mixing context, see the Mixing and Synthesis Tools.

Stereo Effects as Synthesis Tools

Stereo effects deserve dedicated attention in sound design. Stereo is not just a mixing decision — it is a synthesis parameter.

Two slightly detuned oscillators panned hard left and right create a wide, chorused sound that a single oscillator cannot produce. That is stereo synthesis at the oscillator level. But effects can create stereo width from a mono source.

A stereo delay with different delay times on the left and right channels creates a sense of space and movement. A stereo filter with offset cutoff frequencies — left channel slightly higher, right channel slightly lower — creates width through spectral difference. Panning modulated at audio rate (20 Hz and above) creates tremolo effects that blur the boundary between amplitude modulation and spatial movement.

Chorus, at its core, is a short delay with an LFO modulating the delay time, mixed with the dry signal. Stereo chorus uses two delay lines with LFOs running at slightly different rates or phases. The result is a signal that moves in the stereo field while its timbre shifts — effects creating timbral complexity rather than just spatial positioning.

Building a Serial Effects Chain as an Instrument

A powerful exercise: build a serial chain of effects — delays and loopers — and treat the chain itself as a performance instrument. Feed it a drum loop. Include multiple loopers set to different loop lengths, a delay with tempo-synched modulation, and a gating system that crossfades between the processed drums and silence in sync with the looper states.

The performer’s job is not to play drums but to conduct the effects chain — turning loopers on and off, adjusting delay times, modulating feedback in real time. The effects chain is the instrument. The drum loop is raw material.

This is the logical extension of the “effects as synthesis” concept. When you stack enough processing — when feedback feeds into delay feeds into distortion feeds into filtering feeds into more feedback — the output bears no resemblance to the input. The effects chain has become a synthesizer in every meaningful sense. It generates complex, evolving sound from simple input, and the performer controls the timbre through the effect parameters.

What to Practice

• Build a self-oscillating delay in VCV Rack. Set delay time to 3-5 ms, push feedback past unity, and listen to the pitched tone that emerges. Map delay time to a keyboard or sequencer and play it as a melodic instrument.
• Create a flanger from scratch: split a signal, delay one copy by 1-5 ms, modulate the delay time with a slow LFO (0.1-0.5 Hz), mix the copies. Add feedback gradually and listen for the transition from flanging to pitched resonance.
• Patch two oscillators into a ring modulator (a multiplier or bipolar VCA). Start with both at harmonically related frequencies (octave, fifth). Slowly detune one and listen for the inharmonic sum and difference tones.
• Feed a voice or vocal sample through a ring modulator with a fixed sine wave carrier at different frequencies: 100 Hz, 300 Hz, 800 Hz. Notice how each frequency transforms the voice differently.
• Set up a waveshaper or wavefolder and feed it a sine wave. Increase the input gain gradually. Listen to how harmonics appear and change as the drive increases. Follow the waveshaper with a lowpass filter and use the filter to sculpt the added harmonics.
• Build a serial looper chain: feed a simple drum loop through two or three delay/looper modules in series, each set to different loop lengths. Add a VCA gate system so you can crossfade between processed and dry. Perform by manipulating the chain in real time.
• Pick a distorted bass sound from any record you know well. Identify the distortion type (soft clip, hard clip, foldback, bitcrush) and try to match it using VCV Rack modules. The exercise is ear training as much as patching.