Glossary

A 303 line is a sequenced bass line that combines slides with accents, as the built-in sequencer of the Roland TB-303 bass synthesiser of 1981 did.

Step sequencers : Ratchets and slides

A440 is the standard reference pitch that tunes the A above middle C, written A4, to 440 Hz. In equal temperament every other note follows from it, and orchestras sometimes tune higher or lower, for example to 442 Hz or 415 Hz.

Pitch and frequency : Notes, semitones and cents

An AC-coupled input has a capacitor that blocks steady voltages and weakens slow ones. On a modular synthesiser audio inputs are often AC-coupled, so an LFO sent to one arrives much smaller or not at all.

Control signals and audio signals : On a modular synthesiser

An accent marks a step that plays louder than the others. On many drum machines accent is a row of its own and applies to every sound on that step, so it changes which hits stand out without changing which steps play.

Trigger sequencers : Accent Also in Step sequencers

An AD envelope has only attack and decay stages. It runs both to the end as soon as it receives a gate or trigger, whatever the length of the gate, which suits sounds that die away on their own, such as drums and plucked strings.

Envelopes : AD envelopes

Additive synthesis builds a sound by adding sine waves, called partials, and setting the frequency, level and often the envelope of each one separately. Adding all the harmonics at 1/n levels builds a sawtooth, and partials that are not whole-number multiples of the lowest one give metallic and bell-like sounds.

Additive Also in Waveforms , Harmonics and overtones

ADSR stands for attack, decay, sustain and release, the four stages of the most common kind of envelope. Attack, decay and release are times, and sustain is a level held while the gate stays high. The four stages came from a request by the composer Vladimir Ussachevsky to the instrument designer Robert Moog in the mid-1960s.

Envelopes

In FM synthesis an algorithm is a fixed arrangement that sets which operators modulate which and which are heard. The Yamaha DX7 has six operators per voice and 32 algorithms.

FM : Operators and algorithms

An algorithmic reverb builds its reflections from networks of delays. The physicist Manfred Schroeder showed in the early 1960s how to make a natural-sounding digital reverb from a few comb filters and allpass filters, and most algorithmic reverbs since build on that approach.

Reverb : How reverb is made

An allpass filter passes every frequency at the same level but shifts their phase: frequencies well below its centre are shifted only slightly and frequencies above it are shifted further and further out of step. Heard on its own it sounds almost the same as its input, and mixed with the original it produces notches.

Phaser Also in Reverb

Amplitude is how far a signal swings either side of its centre line. A larger amplitude sounds louder. Level is usually measured in decibels, a logarithmic unit that matches the roughly logarithmic way loudness is heard, and doubling the amplitude raises the level by about 6 decibels.

What a signal is : Amplitude Also in The VCA

An amplitude envelope is an envelope connected to a VCA's control input, so that the gate starts the envelope, the envelope opens the VCA, and the oscillator is heard only while the envelope is above zero. On a synthesiser with its own envelope section, the envelope connected this way is labelled amplitude envelope or amp envelope.

The VCA : An envelope into a VCA Also in Envelopes , Envelopes as modulation

Amplitude modulation moves the level of a carrier up and down with a modulator. At audio rate it keeps the carrier and adds two sidebands, at the carrier's frequency plus and minus the modulator's. At a slow rate the sidebands are close enough to the carrier that the ear hears one tone getting louder and quieter.

Modulation at audio rate : Sidebands Also in Ring modulation

AND is a logic rule whose output is high only while both inputs are high. Applied to two rhythms it keeps only the hits they share, so every third step AND every fourth step gives every twelfth step.

Logic : Combining rhythms

Asymmetric clipping clips one half of the wave earlier than the other, as many real circuits, including single valve stages, do. It breaks the wave's half-cycle symmetry and adds even harmonics, whereas a curve that treats both halves alike adds only odd harmonics.

Distortion : Even and odd

Attack is the time an envelope takes to rise from zero to its peak once the gate goes high, set in milliseconds.

Envelopes

On a compressor, attack sets how quickly the gain is turned down once the signal goes over the threshold. An attack of 20 or 30 milliseconds lets the first moment of a drum hit through at full level. An attack of a millisecond or less turns down the peak itself.

Compression : Attack and release

On an envelope follower, attack sets how fast the output rises when the input gets louder. An attack of a few milliseconds lets the output follow each drum hit.

Envelope followers : Attack and release

A signal at audio rate changes faster than about 20 Hz, so it is heard as a tone when listened to. When an audio-rate signal moves a setting, the setting changes too fast to follow, and the result is heard as a change of timbre.

Control signals and audio signals : Speed Also in Modulation at audio rate

Audio-rate modulation is modulation by a signal faster than about 20 Hz. At that speed the setting moves too fast to be heard as movement, and the modulation is heard as a change in timbre because it adds new frequencies, called sidebands, to the sound.

Modulation at audio rate Also in FM , Ring modulation

An auto-wah is a filter whose cutoff is moved by an envelope follower on the input, so each note opens the filter in proportion to how hard it is played.

Envelope followers : Uses

A bandpass filter keeps a band of frequencies around its cutoff and reduces the frequencies on both sides. Narrowed far enough, it leaves a thin band with a vocal, telephone-like sound, and raising the resonance narrows it further towards a single pitch.

Filters : Highpass and bandpass Also in Noise

The base pitch is the pitch a pitch envelope falls to at the end of each hit. Together with the amplifier's decay it sets the low, sustained part of a synthesised drum, and a kick drum's base pitch is usually between about 45 and 60 Hz.

An envelope as a pitch source : Amount and decay

Bessel functions are a set of curves that give the level of each sideband in FM for a given index. Because they rise and fall, some sidebands grow, fall back and grow again as the index increases, and the carrier drops to silence near an index of 2.4. The zeros of Bessel functions also give the mode frequencies of an ideal circular membrane.

FM : Index Also in Physical modelling

A bipolar signal swings either side of zero. A bipolar LFO moves a setting both above and below where it is set, which suits vibrato, where the pitch moves either side of the note.

The LFO : Bipolar and unipolar

Bit reduction lowers the number of levels a signal can take, as in early digital samplers and game consoles, which stored sound with few bits. It turns smooth waves into staircases and produces a gritty, digital kind of distortion.

Distortion : Bit reduction and chords

Blending mixes the two neighbouring frames of a wavetable in proportion to how close the position is to each, which gives a smooth sweep. Stepping instead jumps from one frame to the next, which gives a sweep in audible stages unless the frames are closely spaced.

Wavetables : Blending and stepping

Brightness is the quality of a sound that depends mostly on how quickly the levels of its harmonics fall as the harmonic number rises. Slower rolloff sounds brighter or buzzier, and faster rolloff sounds darker and rounder. A lowpass filter changes brightness by making the levels fall faster above its cutoff.

Harmonics and overtones : Brightness Also in Filters

Brown noise falls in level by 6 dB per octave and sounds like a low roar. It is produced by adding up white noise over time, a random walk, and is named after the botanist Robert Brown, who described the random motion of particles suspended in water in 1827.

Noise : White, pink and brown

A bucket-brigade device is a chain of capacitors that stores a signal and passes it along, which analogue delays use to produce their delay. Roland's Juno-60 and Juno-106 synthesisers used bucket-brigade delays for their chorus.

Delay Also in Chorus

Calibrating an oscillator means adjusting its scale until octaves played from a known voltage source measure as exact doublings of frequency. Analogue oscillators are calibrated with a scale trimmer, and digital oscillators that read 1V/oct are calibrated the same way in software.

One volt per octave : Tracking and calibration

The carrier is the signal whose pitch or level is being modulated. In FM synthesis the carrier is the oscillator heard at the output, and the modulator changes its frequency.

Modulation at audio rate Also in FM , Ring modulation

The ceiling is the level that a limiter keeps the signal's peaks at or below.

Mastering : A mastering chain

A cent is one hundredth of an equal-tempered semitone, which makes 1200 cents to the octave. Cents measure the ratio between two frequencies, so 50 cents is the same size of interval at any pitch.

Pitch and frequency : Notes, semitones and cents Also in One volt per octave

Chorus mixes a signal with a copy delayed by around 15 to 30 milliseconds and moves the delay time slowly with an LFO. The moving delay shifts the copy's pitch a few cents up and down, so the two drift in and out of tune like several players and sound wider and fuller.

Chorus Also in Flanger

A chorus voice is one delayed copy of the signal with its own delay line and LFO. Two or three voices with their LFOs at different phases give a denser ensemble sound, and sending different voices to the left and right speakers widens the sound.

Chorus : Voices and width

The cinquillo is a rhythm of five hits across eight steps spaced 2-1-2-1-2, which is a Euclidean rhythm of five in eight rotated by two steps.

Trigger sequencers : Euclidean rhythms

A click is heard when a sound's level jumps between silence and full level in an instant, because a jump that fast spreads energy across a wide range of frequencies. A gate sent straight to an amplifier, or an envelope with zero attack or release, produces one, and a few milliseconds of attack and release removes it.

Gates and triggers : A gate on its own Also in Envelopes

Clipping flattens the peaks of a signal driven past the largest level a circuit or a digital system can pass, because the output cannot go any further. In a digital system the limit is 0 dBFS.

Distortion Also in Mixing , Mastering , Wavefolding

A clock is a steady stream of short pulses, and each part of a synthesiser connected to it acts once for every pulse. In analogue hardware each pulse is a voltage that jumps up and drops straight back, and over MIDI it is a clock message. A clock pulse marks a moment and does not set how long a note lasts.

Clocks Also in Where pitch comes from , Gates and triggers

A clock divider counts incoming pulses and puts out one pulse for every two, three, four or however many it is set to. Its outputs are counted from the source clock, so they cannot drift away from it.

Clocks : One clock, several rates Also in Logic

A clock multiplier adds pulses between the ones it receives. It measures the gap between the last two incoming pulses and spaces the new ones as though the next gap will be the same, so when the tempo changes the output is out of time for one pulse.

Clocks : One clock, several rates

A comb filter mixes a signal with a short delayed copy of itself. Frequencies whose cycles fit the delay a whole number of times add, and frequencies halfway between cancel, so the response is a row of evenly spaced peaks and notches on a linear frequency axis. A delay of 5 milliseconds gives peaks at 200 Hz and every multiple of 200 Hz.

Comb filters Also in Physical modelling

The delay time of a comb filter sets the spacing of its peaks, which fall at whole-number multiples of one divided by the delay time and form a harmonic series. A shorter delay spaces the peaks further apart and raises their fundamental, so sweeping the delay on noise sounds like a pitch rising and falling.

Comb filters : Delay and pitch

A comparator turns a continuous control signal into a gate: its output is high while the input is above a threshold and low while it is below. Sent an LFO it gives a square wave whose width depends on the threshold, and sent a random signal it gives gates at unpredictable moments.

Logic : Comparators

A compressor turns a signal down when its level goes above a threshold, by an amount set by the ratio. It contains an envelope follower that measures the level and a VCA that sets the gain, and it reduces the difference between the loud and quiet parts.

Compression

Constant-time glide is a glide in which every slide between notes takes the same time whatever the interval.

Slew : Constant rate, constant time and exponential

Contour generator is the name Moog used for the envelopes on the Minimoog, which has one for the filter and another for loudness.

The classic voice : The Minimoog

A control signal is a signal used to move a setting, such as an oscillator's pitch or a filter's cutoff, rather than to be heard. On analogue hardware it is a voltage and is called a control voltage (CV); in software it is a stream of numbers, and over MIDI it is a message. It is the same kind of signal as audio, usually running below about 20 Hz.

Control signals and audio signals : Speed Also in Where pitch comes from , The VCA , Envelopes

Convolution reverb combines a sound with an impulse response of a space, so that every point of the sound is repeated with that space's pattern of reflections. It reproduces the recorded space closely.

Reverb : How reverb is made

A counter is a circuit, built from a chain of flip-flops, that counts the pulses it receives. Counters are the circuits inside clock dividers and sequencers.

Logic : Flip-flops and counting

The cutoff is the frequency at which a filter with no resonance has reduced the signal by 3 decibels, roughly where the reduction starts to be heard. On a lowpass filter, frequencies further above the cutoff are reduced by more, at a rate set by the slope.

Filters : Cutoff and slope Also in Noise , Envelopes

Damping sets how quickly a modelled object loses energy. In the Karplus-Strong string it makes the averaging stronger, so the string loses its high harmonics sooner, as a string muted by the palm does, and in a modal drum it shortens every mode, the higher ones most.

Physical modelling : A string as a delay line

Damping on a reverb makes the tail lose high frequencies faster than low ones as it fades, modelling the way soft surfaces and air absorb high frequencies more than low ones. A heavily damped tail grows darker as it fades. A lightly damped tail, like that of a space with bare stone and glass walls, stays bright.

Reverb : Decay, pre-delay and damping

dBFS is the unit of digital signal level, in decibels measured against the digital maximum of 0 dBFS. Peaks above 0 dBFS are clipped.

Mastering : Loudness and peaks Also in Mixing

A DC-coupled input passes everything from a steady voltage up to audio rate unchanged. On a modular synthesiser control inputs are DC-coupled, which is why cross-patching usually goes into a CV input.

Control signals and audio signals : On a modular synthesiser

DCW, for digital controlled wave, is the control on a Casio CZ synthesiser that sets the amount of phase distortion, from a sine wave at its minimum to the full chosen waveform at its maximum. It has its own eight-stage envelope.

Phase distortion : The DCW envelope

A de-esser is a compressor whose sidechain passes through an equaliser that boosts a chosen range of a vocal's frequencies, so it turns down only the moments when that range is strong.

Compression : Sidechains

Decay is the time an envelope takes to fall from its peak to the sustain level after the attack, set in milliseconds. In an AD envelope the decay falls all the way to zero.

Envelopes

On a reverb, decay sets the reverberation time, which the ear hears as the size of the space. A decay under about half a second sounds like a room, one to two seconds like a hall, and several seconds like a cathedral or a cave.

Reverb : Decay, pre-delay and damping

The decibel is a logarithmic unit of level that matches the roughly logarithmic way loudness is heard. Doubling a signal's amplitude raises its level by about 6 dB.

What a signal is : Amplitude Also in Filters , Noise , The VCA

A delay is an effect that stores a signal and plays it back after a set time. Early units stored it on magnetic tape, analogue delays in a bucket-brigade device, and digital delays as numbers in memory.

Delay Also in Chorus , Flanger

A delay line holds a signal for a set time before passing it on. In a string model, a delay line one period long stands for the time a wave takes to travel along the string and back, and its contents represent the shape of the string at that moment.

Physical modelling : A string as a delay line Also in Comb filters

Delay time is the gap between a signal and its delayed copy, set in milliseconds or as a note length. Below about 30 milliseconds the copy changes the tone of the sound by comb filtering rather than being heard as a separate echo, and above about 150 milliseconds each echo is heard as a separate event.

Delay : Time

Depth is the control that sets how far an LFO moves the setting it is sent to. At zero depth the setting does not move, and at a high depth an LFO on cutoff sweeps the filter across most of its range.

The LFO : Rate and depth Also in Build it in seven steps

On a chorus, flanger or phaser, depth sets how far the LFO moves the delay time or the filters. On a chorus, a higher depth gives a larger pitch shift.

Chorus : A chorus Also in Flanger , Phaser

Detune is a small difference in pitch between copies of a sound. In a chorus the moving delay detunes the copy a few cents either side of the original, and two people playing the same part are detuned from each other in the same way.

Chorus

Digital waveguide synthesis generalises the Karplus-Strong algorithm, with delay lines for each direction of travel and filters for the body and the ends of the instrument. Julius O. Smith, a researcher at Stanford University, developed it in the late 1980s.

Physical modelling : A string as a delay line

The direct sound is the sound that reaches a listener by the shortest path, so it arrives before any reflection from the surfaces of a space.

Reverb

Direction sets the order in which a sequencer reads its steps. Forward plays them in order, backward reverses the pattern, pendulum plays forward then backward, and random chooses any step on each pulse, so the pattern's notes play in an order that changes each time round.

Step sequencers : Length and direction

Distortion changes the shape of a waveform, usually by flattening its peaks when the signal goes past a limit. The flattened peaks add harmonics, so the further the signal is driven past the limit, the brighter it sounds.

Distortion

A dotted eighth is a note length of three quarters of a beat, 450 milliseconds at 100 BPM. As a delay time it places the echoes off the beat, between notes played on the beat.

Delay : Time

A drawbar is a sliding control on a Hammond organ that sets the level of one harmonic of the note, from a tonewheel producing a nearly sinusoidal tone. The drawbars are marked in organ-pipe footages: 8′ is the fundamental, 4′ the second harmonic and 2⅔′ the third.

Additive : Sixteen partials, each with its own level

Drive sets how far past the clipping limit a distortion pushes the signal. As the drive rises, both soft and hard clipping bring a sine wave closer to a square wave, because the sine spends almost all of each cycle flattened at one limit or the other.

Distortion : Soft and hard Also in Mastering , Wavefolding

Ducking lowers the level of one sound whenever another plays, by using the second sound's level to control the first. A pad ducked by a kick drum drops on every kick and swells back between them. It is usually done with a compressor's sidechain input.

Envelope followers : Uses Also in Compression

Dynamic range is the difference between the loudest and quietest moments of a signal. Compression reduces it, and in mastering the gap between peak level and loudness measures how much of it a master keeps.

Compression Also in Mastering

Early reflections are the first few distinct reflections from the nearest walls, the floor and the ceiling, which arrive a few milliseconds after the direct sound. After them, reflections of reflections build up into a dense tail in which single reflections can no longer be heard separately.

Reverb

East coast synthesis is an informal name for the subtractive approach of Robert Moog's instruments, which starts with a waveform that has many harmonics and filters some of them away. It is the counterpart of west coast synthesis.

Wavefolding : West coast

An echo chamber is a hard-walled room that recording studios used to make reverb, by playing sound into it through a speaker and recording it back with a microphone.

Reverb : How reverb is made

An end-of-cycle output sends a trigger when an envelope generator finishes its cycle. Connecting it back to the envelope's own trigger input makes the envelope loop.

Patches that drive themselves : A loop from an envelope

Envelope amount sets how far an envelope moves the setting it is sent to. On a filter, the cutoff control sets the cutoff while the envelope is at zero, and the amount sets how far above that the envelope's peak raises it.

Envelopes as modulation : Amount Also in The classic voice

Most analogue envelope generators move in curves rather than straight lines, because each stage charges or discharges a capacitor through a resistor, which moves fast at first and slows as it nears its target. Loudness is heard on a roughly logarithmic scale, so a curved release sounds like an even fade. A straight-line release sounds loud for most of its length and then drops away quickly.

Envelopes : Straight lines and curves Also in The VCA

An envelope follower turns the level of an audio signal into a control signal. It rectifies the audio and smooths the result with separate attack and release times, so its output rises when the input gets louder and falls when it gets quieter. Envelope followers are also the measuring stage inside compressors, noise gates and limiters.

Envelope followers Also in Compression

An envelope generator produces a control signal that rises and falls in a set shape each time it receives a gate. Connected to a VCA's control input it shapes the level of each note, and connected to a filter's cutoff or an oscillator's pitch it changes those over each note.

Envelopes Also in The VCA , Envelopes as modulation

An equaliser raises or lowers the level of ranges of frequencies to change the tonal balance of a signal. In mastering it usually makes small, broad changes of a decibel or two, because a change to the whole mix affects every sound with energy in that frequency range.

Mastering : Equalising, checking and delivery Also in Mixing

A Euclidean rhythm spreads a number of hits as evenly as possible across a number of steps, for example three hits across eight steps. The pattern can be calculated with the procedure that the ancient Greek mathematician Euclid described for finding the greatest common divisor, and many sequencers generate these rhythms from two numbers and a rotation.

Trigger sequencers : Rhythms from an algorithm

The exponential converter is the part of an analogue oscillator that turns a 1V/oct control voltage into the current that sets the frequency. It is usually built from a matched pair of transistors, whose behaviour changes with temperature, which is why analogue oscillators drift as they warm up and many designs add temperature compensation.

One volt per octave : An exponential conversion

Most synthesisers' pitch controls are exponential: adding a fixed amount to the control raises the frequency by a fixed ratio. The same turn of a knob or the same change in the pitch control signal then gives the same interval at any pitch.

Pitch and frequency : Ratios, not differences Also in One volt per octave

An exponential pitch fall is a fall that is a straight line in cents or octaves, so the frequency halves in equal steps of time. It is heard as an even fall because the ear hears pitch by ratio. A fall that is a straight line in hertz spends most of its time in the upper octaves and drops through the lower ones suddenly at the end.

An envelope as a pitch source : Why the fall is exponential

Exponential slew, also called a lag, moves the output fast at the start and slows as it nears the target, the way a capacitor charges through a resistor. It needs only a resistor and a capacitor, and many analogue portamento circuits work this way.

Slew : Constant rate, constant time and exponential

An exponential VCA maps the control signal to decibels, so equal steps of control give equal steps of loudness as the ear hears it. It suits audio fades and amplitude envelopes.

The VCA : Linear and exponential Also in Envelopes

A fader is a channel's level control. Faders are marked in decibels because loudness is heard on a roughly logarithmic scale, so equal movements give equal changes in loudness.

Mixing : Level and balance

Feedback sends part of a delay's output back into its input, so each echo is delayed again and produces another, quieter echo. At 50% each repeat is half the level of the one before, about 6 dB quieter; close to 100% the repeats barely fade, and above 100% they grow louder each time.

Delay : Feedback and tone

On a flanger, feedback sends part of the output back into the delay, so the peaks between the notches grow taller and narrower and the flanger rings at the pitch set by the delay. Negative feedback inverts the copy, which swaps the positions of the peaks and notches.

Flanger : Feedback and through-zero Also in Phaser

Feedback FM sends an oscillator's output back to its own frequency input. The oscillator speeds up while its output is high and slows down while it is low, which changes a sine's shape towards a ramp and adds harmonics, and with more feedback the output becomes noise. Yamaha's FM synthesisers include feedback on one operator.

Patches that drive themselves : Feedback at audio rate Also in FM

Feedback patching sends the output of a module or other part of a synthesiser, directly or through other parts, back into one of its own inputs. That part's output then depends on its own earlier output, and the patch can keep running and changing with no outside input.

Patches that drive themselves Also in Comb filters

A fifth is the interval with a frequency ratio of 3:2, the ratio between the second and third harmonics. The equal-tempered fifth differs from the 3:2 ratio by about 2 cents.

Pitch and frequency : Ratios, not differences

A filter reduces some of the frequencies in a sound and lets the rest through. In a synthesiser it reduces harmonics to change the brightness of a sound without changing its pitch, and its cutoff is the setting most often modulated in a subtractive voice.

Filters Also in Harmonics and overtones , Noise

A filter envelope is an envelope sent to a filter's cutoff, so that the cutoff rises and falls once for each note. With a fast attack and a short decay to a low sustain, each note starts bright and quickly darkens while its level stays the same.

Envelopes as modulation : Two envelopes, two timescales Also in The classic voice , Build it in seven steps

A filter sweep is a repeating rise and fall in a filter's cutoff, produced by sending an LFO to the cutoff. A rate of one cycle every few seconds gives a slow sweep between a dark and a bright tone, and a few cycles a second give a rhythmic wah.

The LFO : Rate and depth Also in Build it in seven steps

Filtered noise is noise passed through a filter to shape its spectrum. A lowpass leaves a rush, a highpass leaves a hiss used for hi-hats, and a resonant bandpass leaves a narrow band with a vague pitch that sounds like wind when its cutoff is swept slowly.

Noise : Filtering noise Also in Filters

A fixed-frequency modulator stays at the same frequency whatever note is played, so it shifts every frequency of the note up and down by the same number of hertz. A fixed shift in hertz is a different interval at every pitch, so each key gives a differently spaced set of frequencies and the intervals of a melody are not kept.

Ring modulation : Fixed and following modulators

A flanger mixes a signal with a copy delayed by about 1 to 10 milliseconds and sweeps the delay with an LFO. The mix is a comb filter whose notches move up and down the spectrum, which is heard as a sweep. The effect was first made with two tape machines, by pressing on the rim, or flange, of one reel so that its copy fell a few milliseconds behind.

Flanger

A flip-flop is a circuit whose output changes state on each trigger it receives: high on the first, low on the second, high again on the third. Its output is a gate at half the rate of its input, so it divides a clock by two, and a chain of flip-flops divides by four, eight and sixteen.

Logic : Flip-flops and counting

FM synthesis makes sound by changing one oscillator's frequency with another at audio rate. The ratio between the two frequencies sets where the sidebands fall, and the modulation index sets how many of them are strong, usually moved over each note by an envelope. John Chowning, a composer and researcher at Stanford University, found the technique in 1967.

FM Also in Modulation at audio rate

A following modulator runs at a set ratio of the note's frequency. With a whole-number ratio the sums and differences land on harmonics of a common fundamental, so every key gives the same pattern transposed and the result keeps a clear pitch, while ratios that are not whole numbers give a metallic, inharmonic sound.

Ring modulation : Fixed and following modulators Also in FM

A formant is a peak of energy in a sound's spectrum. Vowels are distinguished by where their formants fall, so a wavetable whose frames move a peak from one harmonic to another gives a sweep with a vocal quality.

Wavetables : Frames and position

Four on the floor is a drum pattern with the kick on every beat of the bar, which sets a steady pulse for the other sounds to fill in.

Trigger sequencers : A trigger sequencer

A frame is one single-cycle waveform in a wavetable, so it has the harmonics of that shape. The order of the frames sets how the timbre changes as the position moves, and wavetables commonly have 64 or 256 frames.

Wavetables : Frames and position

A granular instrument freezes a sound by holding the position the grains are taken from still, so one moment of the recording continues for as long as it plays. The scan control sets how fast the position moves through the recording, and a scan of 0 freezes it.

Granular : Position and pitch

Frequency is how many times a signal repeats its shape each second, measured in hertz. People hear repeating signals between roughly 20 Hz and 20 kHz as pitch, and below about 20 Hz the repetition is too slow to fuse into a tone.

What a signal is : Frequency Also in Pitch and frequency , Control signals and audio signals

The frequency deviation is how far the modulator pushes the carrier's frequency away from its unmodulated value in frequency modulation. Divided by the modulator's frequency, it gives the modulation index.

FM : Index Also in Modulation at audio rate

Frequency modulation moves a carrier's pitch with a modulator. At audio rate it produces a series of sidebands at the carrier plus and minus whole multiples of the modulator's frequency, and the modulation index sets how many of them are strong.

Modulation at audio rate : Frequency modulation Also in FM

In FM synthesis the ratio sets the modulator's frequency relative to the carrier's, so the relationship holds across the keyboard. A whole-number ratio puts every sideband on a harmonic of the carrier: a ratio of 1 gives every harmonic and a ratio of 2 gives only the odd harmonics. A ratio that is not a simple fraction, such as 1.41 or 3.14, puts the sidebands between the harmonics and gives an inharmonic, bell-like sound.

FM : Ratio Also in Modulation at audio rate , Ring modulation

A filter's frequency response is how much of each frequency it lets through, drawn as a curve against frequency. Drawn over the harmonics of a source, it shows how much each harmonic is reduced.

Filters

A function generator is a module built around a slew circuit. A gate sent through it gives an attack and release envelope, and with its output fed back to trigger itself it produces a looping envelope.

Slew : Other uses

The fundamental is the lowest partial of a pitched sound, the sine wave at the waveform's repeat rate. It usually sets the pitch you hear, and the harmonics sit at whole-number multiples of its frequency.

Harmonics and overtones Also in Waveforms , Pitch and frequency

Fuzz is clipping so hard that the waveform becomes close to a square wave.

Distortion

Gain reduction is the amount, in decibels, by which a compressor is turning the signal down at each moment.

Compression : A compressor

Gain staging keeps the signal at every stage of the signal path loud enough to stay well above the noise of each stage, and quiet enough to leave headroom below the point where the next stage clips. In a digital mix it mostly means keeping the master's peaks below 0 dBFS.

Mixing : Frequencies and headroom

A gate is a signal that is high for as long as a note should last and low the rest of the time, so it carries both when a note starts and how long it lasts. On a modular synthesiser a gate is usually 5 or 10 volts while high and 0 volts while low, and a keyboard produces one alongside its pitch.

Gates and triggers Also in Envelopes , Where pitch comes from

A gate distributor sends each incoming gate to one of several outputs, so that each note's gate reaches only one voice. Placed after a sequencer's gate output, it is one way to hocket a line on a modular synthesiser.

Hocketing

Gate length is how much of each sequencer step the gate stays high for, usually given as a percentage of the step. At 50% each note sounds for half its step, and near 100% the notes almost join.

Gates and triggers : Gate length Also in Step sequencers

A generative patch keeps producing new material on its own, usually through feedback, random sources and clocks that are not in a simple relation to each other. The player sets up the system and guides it rather than playing each note.

Patches that drive themselves : A sequencer in a loop

Glide, also called portamento, makes the pitch slide from one note to the next over a set time instead of jumping. On some monophonic synthesisers it can be set to happen only when one key is pressed before the previous one is released, and it is usually produced by a slew limiter on the pitch signal.

The classic voice : Monophonic, glide and the keyboard Also in Slew , Step sequencers

A grain is a short sound, typically between 10 and 200 milliseconds long, cut from a recording or generated by an oscillator and given an envelope that fades it in and out. Short grains sound buzzy and dense, and long grains sound smoother and closer to the original recording.

Granular : Grains

Grain density sets how many grains a granular instrument plays per second.

Granular : Grains from a recording

A grain envelope sets how a grain's level rises and falls. A rectangular envelope starts and stops the grain abruptly and adds a click at each edge, and smooth envelopes such as the raised cosine (Hann) and Gaussian shapes fade the grain in and out without a click.

Granular : Grain envelopes

Granular synthesis builds a sound from many short sounds called grains, each shaped by its own envelope and often overlapping. The grains can be cut from a recording, which lets a sound be frozen, stretched or repitched, or generated by an oscillator.

Granular : Grains

Granulation is granular synthesis with grains cut from a recording or a live input, so the result keeps the timbre of the source while its timing, pitch and texture change.

Granular : Synthetic and sampled grains

Hard clipping passes a signal unchanged up to a limit and cuts it off flat above it. The waveform gains sharp corners, which add strong high harmonics.

Distortion : Soft and hard

A harmonic is a partial whose frequency is a whole-number multiple of the fundamental: the second harmonic is at twice its frequency, the third at three times, and so on. Partials at harmonic frequencies are heard together as one note with one pitch.

Harmonics and overtones Also in Waveforms , Filters

The harmonic series is the set of frequencies at whole-number multiples of a fundamental. Strings and air columns vibrate with partials at or very close to this series, which is why they have a clear pitch.

Harmonics and overtones Also in Waveforms , Pitch and frequency

Headroom is the margin between a signal's peaks and the maximum level a stage can carry before it clips, 0 dBFS in a digital system. A mix usually leaves a few decibels of headroom on the master.

Mixing : Frequencies and headroom Also in Mastering

The hertz is the unit of frequency, equal to one cycle per second. A signal at 440 Hz repeats its shape 440 times a second, and a kilohertz (kHz) is 1000 Hz. The unit is named after the physicist Heinrich Hertz.

What a signal is : Frequency Also in Pitch and frequency

A highpass filter reduces the frequencies below its cutoff and keeps the ones above it. It is often used to remove low frequencies from a sound so that it does not overlap a bass part.

Filters : Highpass and bandpass Also in Noise

Hocketing shares a single melody between several voices so that each plays only some of its notes, with the voices taking turns. On a synthesiser the pitch can go to every voice while each note's gate is sent to only one voice, and because each voice has its own envelope, a note keeps ringing through its release while the next voice plays.

Hocketing Also in Switches

Hz/V, or hertz per volt, is a pitch standard in which frequency is proportional to voltage, so doubling the voltage raises the pitch by an octave. A Hz/V instrument and a 1V/oct instrument need a converter between them to play in tune.

One volt per octave : Other standards

An impulse response is a recording of how a space responds to a single click or sweep. Convolution reverb uses it as the pattern of reflections to apply to a sound.

Reverb : How reverb is made

Inharmonic partials are partials at frequencies that are not whole-number multiples of the lowest one. They do not fuse into a single pitch as clearly as harmonics, which gives bells, metal bars, cymbals and drum heads their metallic quality. In a synthesiser they come from FM at ratios that are not whole numbers, from ring modulation and from physical models.

Harmonics and overtones : Inharmonic partials Also in FM , Ring modulation , Physical modelling , Additive

The initial level control on a VCA adds a fixed amount to the control signal. It lets a VCA work as a plain volume control or hold a drone, and with it turned up the VCA never closes completely, so the oscillator is heard between notes.

The VCA : The level knob

An interval is the distance in pitch between two tones, and its size as heard is set by the ratio between their frequencies rather than the difference. An octave is 2:1 at any pitch, while a fixed 100 Hz is more than an octave at the bottom of the piano's range and less than a semitone near the top.

Pitch and frequency : Ratios, not differences

An inverted mapping moves its setting down as the macro moves up. Combining inverted and normal mappings lets one macro brighten a sound while shortening it.

Macro controls : Ranges, directions and curves

ITU-R BS.1770 is the standard that defines loudness measurement in LUFS, with frequency weighting and averaging over time.

Mastering : Loudness and peaks

The jet sound is the pitched, metallic sweep a flanger produces with high feedback, as the comb filter rings at the delay's pitch and that pitch moves up and down.

Flanger : Feedback and through-zero

Just intonation is a tuning system that uses whole-number frequency ratios such as 3:2 and 5:4, taken from the harmonic series. Those two intervals differ from their equal-tempered versions by 2 and 14 cents.

Pitch and frequency : Notes, semitones and cents

The Karplus-Strong algorithm models a plucked string as a delay line one period long, filled with a burst of noise at the pluck. Every sample that leaves the delay line is averaged with its neighbour and fed back in, so the repetition gives the pitch and the averaging, a gentle lowpass filter, makes the high harmonics fade first. Kevin Karplus and Alex Strong, then at Stanford University, published it in 1983.

Physical modelling : A string as a delay line Also in Comb filters

Keyboard tracking raises a filter's cutoff with each higher note played, so that high notes are brighter than low ones.

Filters : What moves the filter

The knee is the region around the threshold where compression begins. With a hard knee, compression starts abruptly at the threshold. With a soft knee, it begins gradually below the threshold and reaches the full ratio above it.

Compression : Threshold and ratio

A lane is a row under the notes of a piano roll that shows one kind of control data over time, such as velocity, pitch bend, a modulation wheel or an automated setting.

Piano rolls : What else a piano roll stores

Legato notes run into each other with no gap between them. On a sequencer, a gate length near 100% produces a legato line.

Gates and triggers : Gate length

An LFO, or low-frequency oscillator, is an oscillator that runs below the audio range, usually from several seconds per cycle up to about 20 Hz. It is sent to a setting and moves that setting up and down: sent to pitch it gives vibrato, sent to a filter's cutoff it gives a repeating sweep, and sent to level it gives tremolo.

The LFO Also in The classic voice , Build it in seven steps , Control signals and audio signals

An LFO delay or fade-in setting holds the LFO off for a moment at the start of each note and then brings it in. It is common on keyboard synthesisers and gives vibrato that arrives part way through a held note.

The LFO : Sync and retrigger Also in Modulating the modulation

A limiter is a compressor with a very high ratio, 10:1 or more, and a fast attack, which stops the signal from going over a set ceiling. Turning up the level into a limiter makes the music louder on average while its peaks stay at the ceiling.

Mastering Also in Compression

Linear slew moves the output at a constant rate in a straight line, so a large step takes longer than a small one: an octave takes twice as long as a tritone.

Slew : Constant rate, constant time and exponential

A linear VCA multiplies its input directly by the control signal, so half the control signal gives half the amplitude. Its arithmetic stays exact, which suits scaling one control signal by another. A linear fade through it sounds loud for most of its length and then drops away quickly.

The VCA : Linear and exponential

A logic module combines gate and trigger signals by treating each one as true while it is high and false while it is low, and produces a new gate according to a rule such as AND, OR, XOR or NOT. In software and on many grooveboxes the same rules appear as conditions on sequencer steps.

Logic

Loop points mark a section of a recording that a sampler repeats, so a short sample can sustain for as long as a key is held. If the end does not join the start smoothly, the join is heard as a click or a bump on every repeat, and placing the points at zero crossings or crossfading across the join are the usual fixes.

Sampling : Loops and direction

A looping envelope is an attack and decay envelope that triggers its own start each time it ends, so it repeats indefinitely at a rate set by the sum of its attack and decay times. It works as an LFO whose shape is set by the two times: a slow rise and fast fall gives a ramp, and equal times give a triangle. Some function generators have a cycle switch that makes this connection internally.

Patches that drive themselves : A loop from an envelope

Loudness normalisation plays each track back at a common loudness by measuring it and turning it up or down, about −14 LUFS on several streaming services. A master pushed louder than that is turned down, so it gains no loudness and keeps the reduced dynamic range that the limiting gave it.

Mastering : The loudness war

The loudness war is the trend through the 1990s and 2000s for masters to be pushed steadily louder by harder limiting, so that the average level rose while the peaks stayed at the ceiling. Drums lost their attack, and quiet passages were brought up close to the level of loud ones.

Mastering : The loudness war

Low cut is the name mixers often give a highpass filter (HPF) on a channel. It removes low frequencies from a part such as a pad or a vocal, which leaves that range clear for the kick drum and bass.

Mixing : Frequencies and headroom

A lowpass filter lets frequencies below its cutoff through and progressively reduces those above it, which makes a bright sound darker without changing its pitch. It is the most common filter type on synthesisers.

Filters : A sawtooth through a filter Also in Harmonics and overtones , Noise

LUFS, loudness units relative to full scale, measures perceived loudness as defined in the ITU-R BS.1770 standard. It weights the signal's frequencies roughly as the ear hears them and averages the result over time, so it tracks how loud music sounds rather than how high its peaks go.

Mastering : Loudness and peaks

A macro control is a single control whose position is sent to several settings at once, each over its own range and in its own direction. Many synthesisers list macros as sources in their modulation matrix. On a modular synthesiser a macro is one control voltage sent through several attenuators to several destinations.

Macro controls Also in The modulation matrix

Makeup gain is gain added after a compressor to bring the overall level back up. With it, the quiet parts end up louder than before and the loud parts about the same, so compressed sounds seem denser.

Compression

A many-to-one switch has several inputs and one output and passes one input at a time. In electronics it is called a multiplexer.

Switches

A mapping turns a macro's position into a value for one setting. Each mapping has a range, which sets where the setting sits at either end of the macro's travel, and a direction, which sets whether the setting rises or falls as the macro rises.

Macro controls : Ranges, directions and curves

A mapping curve sets how a setting changes across the macro's travel, so that it can change mostly at one end. A macro mapped to cutoff often uses an exponential curve, so that each part of the macro's travel changes the brightness by a similar amount.

Macro controls : Ranges, directions and curves

Masking happens when two parts with energy in the same frequency range are heard together and the louder hides the quieter. In a mix it is strongest in the low end, where the kick, the bass and the lowest notes of chords overlap.

Mixing : Frequencies and headroom

Mastering is the last stage of production before release. It treats a finished mix as a single stereo signal and adjusts its tonal balance with equalisation, its dynamics with compression and limiting, and its loudness. It also prepares the files at the sample rate, bit depth and format that each destination requires.

Mastering

Middle C is about 261.63 Hz in twelve-tone equal temperament with A4 at 440 Hz. It is called C4 on this site and by most synthesisers, although some manufacturers call it C3.

Pitch and frequency : Notes, semitones and cents Also in One volt per octave

MIDI clock carries tempo from one instrument to another over a MIDI cable as a stream of clock messages, at 24 pulses per quarter note.

Clocks

A MIDI-to-CV interface turns MIDI messages into the signals a modular synthesiser uses: a pitch signal, a gate and further control voltages.

Piano rolls : What else a piano roll stores

The mix control on an effect sets the balance between the original signal and the processed signal. On a reverb it sets how much reverb is heard, and on a phaser a mix of zero removes the effect, because the allpass filters on their own change only phase.

Reverb : A reverb Also in Delay , Chorus , Phaser

A mix envelope is a stored path that the mix point of a vector synthesiser follows over each note, so the balance of the four sources changes without a hand on the joystick.

Vector : Mix envelopes

A mixer gives each part a channel with controls for its level, its position between the speakers and often a filter, equaliser and effect sends, and adds the channels together into a stereo output. It may be a mixing desk, a modular mixer module or the mixer of a digital audio workstation.

Mixing

Mixing sets the level, stereo position and tone of each part of a piece so that they can all be heard together, and adds effects such as compression and reverb where they help.

Mixing Also in Mastering

Modal synthesis models a vibrating object as a set of modes and plays one decaying sine wave for each, with the model setting how strongly each one starts and how quickly it decays. It suits bars, bells, plates and drum heads, whose modes are known or can be measured.

Physical modelling : Modes

A mode is one of the patterns in which an object such as a drum head, a bell or a metal bar vibrates, each at its own frequency. Any strike sets several modes moving at once. The modes of an ideal circular membrane are at 1, 1.59, 2.14, 2.30 and more times the lowest, none a whole-number multiple, which is why a drum has a less definite pitch than a string.

Physical modelling : Modes

A modular synthesiser is made of separate modules that are connected with patch cables, and no two modules are connected until the player patches them. The modular synthesisers of the 1960s were large cabinets of such modules.

The classic voice : The Minimoog Also in Build it in seven steps

A modulated delay is a delay line whose time is moved continuously, usually by an LFO. While the delay time grows, playback runs slower than real time and the pitch drops. While it shrinks, the pitch rises. The size of the shift depends on how fast the delay time changes rather than on its length.

Chorus : Why a moving delay changes pitch Also in Flanger

Modulating the modulation means using one modulation source to change the depth, rate or shape of another. An envelope on an LFO's depth makes vibrato fade in over each note, and each layer can itself be modulated again so that a patch changes over minutes without the player touching a control.

Modulating the modulation Also in The modulation matrix

The modulation amount sets how strongly a source moves its destination, and it can be positive, negative or zero. A negative amount inverts the source, so an envelope that raises a filter's cutoff with a positive amount lowers it with a negative one. On a modular synthesiser the amount is set with an attenuator or a VCA on each connection.

The modulation matrix : Adding and inverting Also in Modulating the modulation

The depth of a modulation sets how far it moves its destination. Modulating the depth makes the effect grow and shrink, as in vibrato that fades in or a filter sweep that widens over a phrase. On a modular synthesiser depth is set by a VCA between the source and the destination, whose control input can take any other signal.

Modulating the modulation : Depth, rate and shape Also in The modulation matrix , The VCA

A modulation destination is any setting a synthesiser exposes to modulation, such as pitch, cutoff, resonance, level, pan, pulse width or wavetable position. When two sources reach the same destination, their contributions add, so the destination moves by the sum of the two at every moment.

The modulation matrix : Sources and destinations

The modulation index is the frequency deviation divided by the modulator's frequency. At index 0 the carrier is a plain sine, and as the index rises energy moves from the carrier into more and more sidebands, with the level of each following Bessel functions. An envelope on the index changes the brightness over each note without a filter.

FM : Index Also in Modulation at audio rate

A modulation matrix is a table that connects modulation sources to destinations, with an amount for each connection. Some synthesisers show it as a grid with sources down one side and destinations across the top, and others as a list of slots. It does the job of patch cables on a synthesiser with fixed internal connections, and lets any source it lists reach any destination it lists.

The modulation matrix Also in Modulating the modulation , Macro controls , Control signals and audio signals

A modulation source is anything that produces a control signal, such as an LFO, an envelope, a random generator, or one of the player's own controls such as key velocity, the modulation wheel or aftertouch. In a modulation matrix the sources are listed down one side.

The modulation matrix : Sources and destinations

The modulator is the signal that moves the carrier's pitch or level, or multiplies it in ring modulation. Its frequency sets the spacing of the sidebands either side of the carrier.

Modulation at audio rate Also in FM , Ring modulation

A monophonic synthesiser plays one note at a time. Pressing a second key while holding the first moves the same oscillator to the new pitch without starting a new gate, so the envelopes carry on instead of restarting, and letting go of the second key returns to the first if it is still held.

The classic voice : Monophonic, glide and the keyboard

Multisampling records an instrument at several pitches and uses the nearest recording for each key. It keeps each recording close to its original pitch, so the change in character from changing playback speed stays small.

Sampling : Speed and pitch

Mute silences one channel and leaves the rest playing.

Mixing

A negative envelope amount inverts the envelope, so the destination moves down when a note starts and returns as the envelope decays. On a filter it gives notes that start dull and brighten, and it needs the cutoff set high enough for the envelope to lower it.

Envelopes as modulation : Amount

In a comb filter, negative feedback inverts the delayed copy, which swaps the peaks and notches. The peaks move to the odd multiples of half the original fundamental, so the pitch drops an octave and only odd harmonics remain.

Comb filters : Delay and pitch

A node is a point or line on a vibrating object that does not move in a particular mode. A strike or pluck on a node cannot excite that mode, which is why a drum struck at the centre, where every mode with a diameter node is still, sounds rounder and more hollow.

Physical modelling : Modes

Noise is a signal that never repeats. Its energy is spread across a wide range of frequencies in a continuous spectrum, at levels that change from moment to moment at random, so it has no clear pitch.

Noise Also in What a signal is

NOT is a logic rule that inverts a single input, so its output is high wherever the input is low. Applied to a rhythm it turns rests into hits and hits into rests, so NOT on a kick pattern gives a part that fills the gaps around the kick.

Logic : Combining rhythms

A notch is a narrow dip in a response at a frequency where a signal and its altered copy are half a cycle out of step and cancel. In a flanger the first notch is at one divided by twice the delay and the rest at odd multiples of it, so a 1 millisecond delay puts notches at 500 Hz and 1500 Hz.

Flanger : Notches that move Also in Phaser

A note-off is the MIDI message that ends a note. In a piano roll it falls at the end of each note bar.

Piano rolls

A note-on is the MIDI message that starts a note, carrying its key number and velocity. In a piano roll, each note bar is a note-on at its start and a note-off at its end.

Piano rolls

An octave is the interval between two pitches where one frequency is twice the other, such as 220 Hz and 440 Hz. Each octave covers twice as many hertz as the one below it but sounds the same size, and pitches an octave apart are given the same note name in most musical traditions.

Pitch and frequency Also in One volt per octave , Noise

Ohm's acoustic law is the observation, described by the physicist Georg Ohm in 1843, that for a steady tone the ear responds mainly to the level of each harmonic and only slightly to their phases. Two waveforms with the same spectrum and different phases sound almost the same.

Harmonics and overtones : Phase

A one-to-many switch has one input and several outputs and sends the signal to one output at a time. In electronics it is called a demultiplexer.

Switches Also in Hocketing

An operator is a sine oscillator with its own envelope in an FM synthesiser. It can act as a carrier, heard at the output, or as a modulator of another operator, and one operator can modulate itself through feedback.

FM : Operators and algorithms Also in Patches that drive themselves

OR is a logic rule whose output is high while either input is high. Applied to two rhythms it merges them into one busier pattern.

Logic : Combining rhythms

An oscillator is the part of a synthesiser that generates sound by repeating one short pattern, the waveform, hundreds or thousands of times a second. How often the pattern repeats sets the pitch, and its shape sets which harmonics the sound contains.

Waveforms Also in Where pitch comes from , One volt per octave

Overdrive is gentle, soft clipping, the mildest of the names for driving a waveform past the level a circuit can pass cleanly. Distortion is the name for more clipping than that, and fuzz for the most.

Distortion

An overtone is any partial above the fundamental. Overtones are counted from the first one above the fundamental, so the second harmonic is the first overtone.

Harmonics and overtones

Panning sets how much of a channel goes to each speaker, which places the part somewhere between left and right. Central parts such as the kick, the bass and a lead vocal are usually kept in the middle, and parts in the same range can be moved to opposite sides so that each is easier to hear.

Mixing : Panning

A partial is any one of the sine waves that make up a sound. The lowest partial is the fundamental, and partials at whole-number multiples of it are harmonics.

Harmonics and overtones Also in Waveforms , Additive

Patching is connecting the parts of a modular synthesiser with patch cables to build a voice or other system. A newly patched module changes the sound only once its output is patched to an input that affects the sound.

Build it in seven steps Also in The classic voice

Pattern length sets how many steps a sequencer plays before it starts again. A length that does not divide evenly into the bar, such as five or seven steps against four beats, starts again at a different point in the bar each time round, so a five-step pattern lines up with a four-step drum pattern again only every twenty steps.

Step sequencers : Length and direction Also in Trigger sequencers

The peak of a signal is the distance from its centre line to its highest point. Audio swinging from -5 to +5 volts is 5 volts peak.

What a signal is : Amplitude

Peak level, measured in dBFS, is how close the highest samples of a signal come to the digital maximum of 0 dBFS. A peak above 0 dBFS is clipped, but peak level gives little indication of how loud the music sounds.

Mastering : Loudness and peaks

Peak-to-peak is the distance from the lowest point of a signal to its highest point. Audio on a Eurorack modular synthesiser typically swings from -5 to +5 volts, which is 10 volts peak-to-peak.

What a signal is : Amplitude

Pendulum is a sequencer direction that plays the pattern forward and then backward without repeating the end steps.

Step sequencers : Length and direction

The period is the length of time one cycle of a repeating signal lasts. It is the reciprocal of the frequency, so a signal at 220 Hz has a period of 1/220 of a second, about 4.5 milliseconds.

What a signal is : Frequency

The phase of a harmonic is the point in its cycle where it starts. Moving the phases of a tone's harmonics relative to each other changes the waveform's shape a great deal but changes the sound of a steady tone very little.

Harmonics and overtones : Phase

Phase distortion makes a waveform by reading a cosine wave at an uneven rate within each cycle, while keeping the length of the cycle the same. An even rate gives a sine wave, and bending the rate gives shapes close to a sawtooth, a square or a pulse. Casio introduced it in the CZ-101 in 1984.

Phase distortion Also in FM

Phase modulation is the method many synthesisers use to produce FM synthesis. It modulates the oscillator's phase rather than its frequency, and it produces equivalent spectra.

FM : Operators and algorithms Also in Phase distortion

A phase shift moves a frequency out of step with the original by part of a cycle without changing its level. When a signal is mixed with a copy in which a frequency has been shifted by exactly half a cycle, that frequency cancels.

Phaser

A phaser passes a signal through a chain of allpass filters, mixes the result with the original, and sweeps the filters with an LFO. The mix has a few notches where the two cancel, and they are unevenly spaced, unlike the many evenly spaced notches of a flanger.

Phaser Also in Flanger

Physical modelling recreates the sound of an object such as a string, a tube or a drum head by calculating how it vibrates. Its controls are physical properties, such as length, stiffness, damping and where the object is struck or plucked, and the waveform and spectrum come out of the calculation.

Physical modelling

A piano roll is a note editor that shows notes as bars on a grid, with pitch on the vertical axis and time on the horizontal axis. Each bar's position sets when its note starts and its length sets how long the note lasts, and notes can start at any time and overlap, so a piano roll can hold chords. It is the main note editor in most digital audio workstations, and its name comes from the punched paper rolls that drove player pianos.

Piano rolls Also in Step sequencers

A ping-pong delay is a stereo delay that feeds each repeat into the opposite channel's delay, so the echoes alternate between the left and right speakers.

Delay : Stereo

Pink noise has the same power in every octave, so its level falls by 3 dB each time the frequency doubles. It sounds roughly even from bass to treble and is used to measure loudspeakers and rooms.

Noise : White, pink and brown

Pitch is how high or low a tone is heard, and it follows the frequency of the tone's fundamental. The ear compares frequencies by their ratio, so pitch is heard on a logarithmic scale.

Pitch and frequency Also in What a signal is , One volt per octave

The pitch amount sets how far above the base pitch a pitch envelope starts each note, measured in semitones or octaves. A synthesised kick drum typically starts two or three octaves above its base pitch.

An envelope as a pitch source : Amount and decay

The pitch decay sets how quickly a pitch envelope falls back to the base pitch. A decay of 30 to 80 milliseconds with a large amount is heard as a punch, around 100 to 200 milliseconds the fall is heard as a swoop, and longer decays give a sweep or a laser zap.

An envelope as a pitch source : Amount and decay

A pitch envelope is an envelope sent to an oscillator's pitch, which moves the pitch once for each note and returns it to the note's pitch as the envelope decays. A small, slow fall gives a scoop into the note, and a large, fast fall is how most synthesised kick drums and toms are made.

An envelope as a pitch source Also in Envelopes as modulation

A plate reverb makes reverb by vibrating a large steel sheet with the signal and picking up the vibration.

Reverb : How reverb is made

A classic sampler changes pitch by changing playback speed. Playing a recording twice as fast doubles every frequency, raises the pitch an octave and halves the length, and each semitone is a speed change of the twelfth root of two, about 6%.

Sampling : Speed and pitch

The pluck point sets where along a modelled string it is plucked, and it removes the harmonics that have a node at that point. A string plucked at a quarter of its length has no fourth, eighth or twelfth harmonic.

Physical modelling : A string as a delay line

A pole is a stage of a filter that adds 6 decibels per octave to its slope. Filters are often built from two-pole sections with a slope of 12 dB per octave, as the filter in this guide is: two sections in a row give 24 dB per octave and four give 48.

Filters : What resonance does

A polyphonic synthesiser has several voices and shares notes between them so that several notes can sound at once, which lets it play chords. Its voices are usually identical.

Hocketing : Overlapping releases Also in The classic voice

A polyrhythm is two rhythms with different divisions played together, such as clock outputs divided by three and by four. Those two land together on the first beat and again after twelve beats, the smallest number that both three and four divide into.

Clocks : One clock, several rates

A power chord is a fifth, the interval distorted guitar parts often play. Distorting a fifth produces a difference frequency an octave below the lower note and sums that are harmonics of that same frequency, so the result keeps one clear pitch, while denser chords produce frequencies that fall between harmonics.

Distortion : Bit reduction and chords

PPQN, pulses per quarter note, is the number of clock pulses in one beat. Modular sequencers that move one step per pulse are usually clocked at 4 PPQN, which makes each step a sixteenth note, and MIDI clock uses 24.

Clocks

Pre-delay is the gap between the direct sound and the start of the reverb. In a large space the first reflections arrive tens of milliseconds after the direct sound, and adding that gap keeps the start of each note clear of its reverb.

Reverb : Decay, pre-delay and damping

Pumping is the audible rise of a sound's level between loud hits, produced by a compressor with a fast release or by sidechain ducking. It is common in dance music.

Compression : Attack and release Also in Envelope followers

Quantizing moves each note's start time to the nearest line of a chosen grid, such as sixteenth notes, so that notes played slightly early or late start exactly on the grid. Most software can also quantize partially, moving notes only part of the way towards the grid.

Piano rolls : The grid and quantizing

A quantizer moves each value of a continuous pitch signal to the nearest note of a chosen scale, or of any other set of pitches given to it. It rounds each value and has no effect on which value comes next, so a random line stays random and every note belongs to the scale.

Where pitch comes from : Quantizing Also in One volt per octave

A ramp is an LFO shape that moves the setting steadily in one direction through each cycle and then jumps back. A rising ramp moves the setting up and drops it back, and a falling ramp does the opposite, which on cutoff or level gives a repeated swell or a repeated pluck.

The LFO : Shape

The random shape on an LFO jumps to a new random value at the start of every cycle and holds it until the next, which produces the same stepped signal that a sample and hold produces from noise.

The LFO : Shape

A random walk is a signal produced by adding up random values over time, so that each value starts from where the last one ended. Adding up white noise this way produces brown noise.

Noise : White, pink and brown

People hear repeating signals from roughly 20 Hz to 20 kHz as pitch, with the upper limit falling with age. Below about 20 Hz a repeating signal is heard as movement or as separate clicks rather than as a tone.

What a signal is : Frequency Also in Control signals and audio signals

A ratchet divides a sequencer step into several equal gates, so a single step is heard as a fast repeated note or a drum roll. A ratchet of 3 fits three gates into the length of one step.

Step sequencers : Ratchets and slides

Rate is the frequency of an LFO, in hertz or cycles per second. A rate of one cycle every few seconds on a filter's cutoff gives a slow sweep, and above about 20 Hz the movement is too fast to follow and is heard as a change of timbre instead.

The LFO : Rate and depth Also in Modulating the modulation

On a chorus, flanger or phaser, rate sets how fast the LFO moves the delay time or the filters, in hertz. A fast rate on a chorus makes the pitch movement audible as vibrato.

Chorus : A chorus Also in Flanger , Phaser

The ratio sets how strongly a compressor turns down a signal above the threshold. At 4:1, every 4 dB the input rises above the threshold becomes 1 dB at the output, so a signal 8 dB over comes out 2 dB over.

Compression : Threshold and ratio

Rectification makes the negative half of each cycle of a signal count the same as the positive half, so the result is always positive. An envelope follower rectifies audio before smoothing it.

Envelope followers

Release is the time an envelope takes to fall back to zero once the gate goes low, set in milliseconds. If the gate ends before the envelope reaches its peak, the release starts from whatever level the envelope had reached.

Envelopes

On a compressor, release sets how quickly the gain recovers once the signal falls below the threshold. A fast release lets the level rise quickly between hits. A slow release keeps the gain lowered across several hits and can hold down quieter sounds that follow a loud one.

Compression : Attack and release

On an envelope follower, release sets how fast the output falls when the input gets quieter. With a very short attack and release the output follows almost every cycle of the waveform. With a release of half a second or more the hits merge into a smooth line.

Envelope followers : Attack and release

A pulse at a reset input sends a sequencer or clock divider back to its first step. On a modular synthesiser, patching one reset signal to every module brings them all back to step one at the same moment, which brings modules that have fallen out of step back into line.

Clocks : Reset Also in Switches

Resonance is a boost to the frequencies around a filter's cutoff, produced by feedback in the filter. At low settings it makes a sweep audible as a moving peak, and at higher settings the peak becomes narrow and loud enough to pick out individual harmonics as the cutoff passes them.

Filters : What resonance does Also in Noise

A resonant waveform on a phase distortion synthesiser is a sine wave at a higher frequency than the note, restarted at every cycle of the note and faded to zero by a falling ramp before each restart. Raising its frequency moves a peak up through the harmonics, which sounds like sweeping a resonant filter.

Phase distortion : Resonance

A resonator is a comb filter with feedback close to 100%, so it keeps ringing at the fundamental set by its delay and at that fundamental's harmonics after the input stops. Fed a short burst of noise, it rings like a plucked string and fades because a little energy is lost each time round the loop.

Comb filters : Feedback and resonance Also in Physical modelling

A rest on a step sequencer is a step with its gate off. The step still lasts as long as the others, and its pitch has no audible effect because no gate starts an envelope.

Step sequencers : Pitch and gate

Resynthesis measures the frequency and level of each partial in a recording over time and plays the measurements back as sine waves. Changing the measurements before playback changes the sound in ways a filter cannot, such as stretching every partial's frequency.

Additive : Resynthesis

Retrigger restarts an LFO's cycle with each new note, so that every note starts at the same point in the cycle. Without it, a free-running LFO is at whatever point in its cycle it happens to be when the note starts.

The LFO : Sync and retrigger

Reverb is the dense pattern of reflections a space adds to a sound: the direct sound, then a few distinct early reflections, then a dense tail of reflections that fades as each one loses energy to the surfaces it meets. A reverb effect recreates that pattern.

Reverb

Reversing a recording plays it backwards, which turns each sound's shape around in time. A sound that started sharply and decayed then swells and stops abruptly.

Sampling : Loops and direction

A ring modulator multiplies two signals, so its output contains the sums and differences of their frequencies and not the original frequencies. Ring-modulating 500 Hz with 200 Hz gives 300 Hz and 700 Hz. The name comes from the ring of four diodes in the original analogue circuit.

Ring modulation Also in Modulation at audio rate

Rise and fall are separate slew limiter controls that set how quickly the output follows rising and falling changes. With a long rise and a short fall, the output climbs slowly and drops at once, so a melody slides up into high notes and jumps down to low ones.

Slew : Rise and fall

The rising edge is the moment a gate or trigger goes from low to high. A gate begins with the same rising edge a trigger has, so most parts of a synthesiser that act only at the start of an event respond to either.

Gates and triggers : Gate or trigger

Rolloff is the rate at which the levels of a sound's harmonics fall with harmonic number. In a sawtooth the level of harmonic n is 1/n, and in a triangle wave it is 1/n².

Harmonics and overtones : Brightness Also in Waveforms

Rotation moves the starting point of a pattern, such as a Euclidean rhythm, while keeping the spacing of its hits. Rotating a pattern shifts it against the beat.

Trigger sequencers : Euclidean rhythms

Round robin shares notes between voices in a fixed order, so with three voices each voice plays every third note. An eight-note sequence shared between three voices this way repeats its pattern of voices only every twenty-four notes.

Hocketing : Ways to share the notes

RT60 is the reverberation time: the time it takes the reflections in a space to fall by 60 decibels, about one thousandth of their starting level, after the sound stops. A small room might have an RT60 under half a second and a large concert hall around two seconds. The physicist Wallace Clement Sabine measured it at Harvard in the 1890s.

Reverb

A sample and hold takes the value at its input each time it receives a clock pulse and holds that value steady until the next pulse. Fed with noise, it produces a new random value on every tick, which is a common way to make unpredictable melodic lines.

Where pitch comes from : Sample and hold Also in Noise , Clocks

A sampler is an instrument that plays recordings, called samples, back. It can change a recording's playback speed to play it at different pitches, set where playback starts and ends, loop part of it so a note can be held, play it backwards, and shape it with filters and envelopes.

Sampling

A sawtooth wave contains every harmonic of its fundamental, with the level of harmonic n at 1/n of the fundamental. It is the usual starting point for subtractive synthesis.

Waveforms : What each waveform contains Also in Harmonics and overtones , Filters

A scale trimmer is an adjustment, usually on an analogue oscillator's circuit board, that sets how much the frequency changes for each volt at the pitch input. It is used to calibrate the oscillator to exactly one octave per volt.

One volt per octave : Tracking and calibration

A self-clocking sequencer sends its pitch control signal to the rate of the clock that drives it, so each step's pitch sets how long that step lasts. The rhythm then follows the shape of the melody, and changing a note changes the rhythm as well as the tune.

Patches that drive themselves : A sequencer in a loop

Self-oscillation happens when a filter's resonance is high enough that the feedback producing the peak becomes self-sustaining. The filter then produces a sine wave at its cutoff frequency with no input, and it can be played as a sound source.

Filters : What resonance does

A semitone is one step of twelve-tone equal temperament, which multiplies the frequency by the twelfth root of two, about 1.0595. Twelve semitones make an octave, and under 1V/oct a semitone is one twelfth of a volt.

Pitch and frequency : Notes, semitones and cents Also in One volt per octave

A sequencer plays a pattern of steps, each holding a note, and moves to the next step on each clock pulse. Each step sends two signals: a pitch that sets the note and a gate that sets when the note starts and how long it lasts.

Where pitch comes from : A sequencer Also in Clocks , Gates and triggers , Step sequencers

A sequential switch passes one of several inputs to its output, or one input to one of several outputs, and moves to the next position each time it receives a trigger. With a fixed voltage at each input it works as a step sequencer, and many early analogue sequencers were built this way.

Switches : Sequential and voltage-controlled Also in Hocketing

A shelving equaliser raises or lowers the low end or the high end of a signal as a whole, adding or removing bass or treble across the whole mix. Mastering usually uses gentle shelves of a decibel or two.

Mastering : Equalising, checking and delivery

A sideband is a new frequency created by modulation, spaced either side of the modulated frequency by the modulator's frequency or its multiples. Amplitude modulation of a 440 Hz tone at 100 Hz adds sidebands at 340 Hz and 540 Hz, and frequency modulation adds a series of them at the carrier plus and minus one, two, three and more times the modulator's frequency.

Modulation at audio rate : Sidebands Also in FM , Ring modulation

A sidechain is the input a compressor measures when that is a different signal from the one it turns down. A pad compressed with a sidechain from the kick drum drops on every kick.

Compression : Sidechains Also in Envelope followers

A signal is a level that changes over time. In analogue hardware the level is a voltage, and in software it is a stream of numbers. A repeating signal in a synthesiser, whether heard or used for control, is described by its amplitude, frequency and waveshape.

What a signal is Also in Control signals and audio signals

The signal path is the route the audio takes from the oscillator to the output. In a classic subtractive voice it runs from the oscillator to the filter to the amplifier, and the keyboard, envelopes and LFO sit outside it, producing control signals that change settings along it.

The classic voice Also in Build it in seven steps

A sine wave is the waveform that contains only its fundamental, with no harmonics above it. It sounds like a single pure tone, and every other waveform can be built by adding sine waves together.

Waveforms : What each waveform contains

A single-cycle waveform is a recording or drawing of exactly one cycle of a wave. Played repeatedly, it produces a steady tone whose pitch is set by the playback rate and whose timbre is set by the shape of the cycle.

Wavetables

Slapback is a single echo about 50 to 150 milliseconds after the original, heard as a doubling of the sound. It was used on many early rock and roll vocals.

Delay : Time

A slew limiter is a processor that limits how fast a signal can change. A small step at its input passes almost unchanged and a large one comes out as a ramp, so a stepped pitch line comes out as notes that glide into each other. On keyboard synthesisers the same circuit is placed on the pitch signal to produce glide.

Slew Also in The classic voice

Slew rate is the speed at which a signal can change. A slew limiter sets a maximum slew rate for the signal passing through it.

Slew

A slide makes a sequencer step glide into its pitch from the previous note. The previous gate is held high into the sliding step, so the envelopes do not restart and the two notes join.

Step sequencers : Ratchets and slides Also in Slew

A filter's slope is how steeply it reduces frequencies past the cutoff, measured in decibels per octave. At 24 dB per octave a harmonic one octave above the cutoff is reduced by 24 dB, and at 12 dB per octave the same harmonic is reduced by 12 dB and stays clearly audible.

Filters : Cutoff and slope

A slot is one connection in a modulation matrix that is shown as a list, and each slot holds a source, a destination and an amount. On many matrices a slot's destination can be the amount of another slot, which lets one source control how strongly another acts.

The modulation matrix Also in Modulating the modulation

Snap sets the grid lines that a new note's start lands on when it is placed in a piano roll, such as beats, eighths or sixteenths. With snap turned off, a note can start anywhere.

Piano rolls : The grid and quantizing

Soft clipping begins to flatten a waveform before it reaches the limit and approaches the limit gradually. The corners are rounded and the high harmonics are weaker than with hard clipping, and it is often described as warmer.

Distortion : Soft and hard

Solo plays one channel on its own and silences the others, so the channel can be listened to and adjusted without the rest of the mix.

Mixing : Level and balance

A spectrum plots a sound against frequency rather than against time. For a pitched tone each line or bar is one harmonic and its height is that harmonic's level, while noise has a continuous spectrum with every frequency in its range present.

Waveforms : What each waveform contains Also in Harmonics and overtones , Noise

Spray scatters the grains around the set position in the recording instead of taking them all from one point. Grains from slightly different moments then overlap, and the sound at that position becomes a blurred texture.

Granular : Position and pitch

A spring reverb sends the signal down a coiled spring to produce reverb. It is common in guitar amplifiers.

Reverb : How reverb is made

A square wave contains only the odd harmonics, with levels falling as 1/n. The absence of even harmonics is heard as a hollow tone. The second half of each cycle is an upside-down copy of the first, and that symmetry cancels the even harmonics.

Waveforms : What each waveform contains Also in Harmonics and overtones

Staccato notes are short and separated by silence. On a sequencer, a short gate length produces staccato notes while the tempo and pitches stay the same.

Gates and triggers : Gate length

A stage is one first-order allpass filter in a phaser, which can shift the phase by up to half a cycle. Two stages together reach a half-cycle shift at one frequency, so every two stages make one notch: a four-stage phaser has two notches.

Phaser : Stages and notches

The start and end points of a sampler set which part of a recording plays. When looping is on, playback repeats the section between them.

Sampling : A sampler

A step is one of the equal divisions of a sequencer's pattern, holding the values for one position, such as a pitch and a gate. Each clock pulse moves the sequencer on by one step, which is usually a sixteenth note.

Step sequencers Also in Trigger sequencers

A step sequencer divides a pattern into equal steps and stores a pitch and a gate for each one. On every clock pulse it moves to the next step and sends that step's pitch and gate to the voice.

Step sequencers Also in Switches , Piano rolls

Stereo width is how far a sound spreads between the left and right speakers. Sending different delayed copies to each speaker widens a sound, because each ear hears a slightly different mixture.

Chorus : Voices and width Also in Delay

Stretched tuning is the practice of setting a piano's top and bottom octaves to match the partials of its strings rather than exact harmonic positions. The stiffness of a real string raises each partial a little above its harmonic position, by more for higher partials.

Harmonics and overtones : Inharmonic partials

The strike point sets where a modelled drum head is struck, and so how strongly each mode starts. A strike at the centre excites only the modes that rise and fall in rings, and moving it towards the edge brings in the others and makes the sound brighter.

Physical modelling : Modes

Subtractive synthesis starts from an oscillator with many harmonics and shapes the sound mainly by filtering some of them away. The standard layout is an oscillator, then a filter, then an amplifier, with envelopes and an LFO controlling them.

The classic voice Also in Build it in seven steps , Filters

Sum and difference frequencies are the frequencies a ring modulator produces: for each frequency in one input and each in the other, the output contains their sum and their difference. A sawtooth note therefore produces twice as many components as it contains, none of them at the note's own frequencies.

Ring modulation Also in Modulation at audio rate , Distortion

Sustain is the level an envelope holds for as long as the gate stays high, given as a fraction of the peak. It is a level and has no length of its own: the gate sets how long it lasts.

Envelopes Also in Gates and triggers

Swing moves every second clock pulse later and leaves the first pulse of each pair where it was, so the beat does not drift. The setting is the position of the late pulse within the pair: at 50% the rhythm is straight, and at 66.7% each pair becomes the first and last notes of a triplet.

Clocks : Swing Also in Trigger sequencers

On a wavefolder, symmetry adds an offset before the folder so the positive and negative halves of the cycle fold different numbers of times. A sine folded evenly about its centre contains only odd harmonics, and the offset makes the two halves differ so that even harmonics appear.

Wavefolding : Symmetry

Synthetic granular synthesis generates each grain with an oscillator instead of cutting it from a recording, so each grain's frequency, waveform and envelope can be set separately. Short grains at regular times produce a pitch at the grain rate, with a formant at the frequency inside each grain.

Granular : Synthetic and sampled grains

A tape delay records a signal onto a loop of magnetic tape and plays it back from a head further along the tape, so the delay time is set by the distance between the heads and the speed of the tape. Each pass through the tape loses some high frequencies, so with feedback the later echoes are progressively duller.

Delay

Tempo is the speed of the beat, normally given in beats per minute (BPM). Dividing by 60 gives beats per second, which is a frequency in hertz, so 120 BPM is 2 Hz.

Clocks

Tempo sync sets a rate or a time as a note length, such as one cycle per bar or an eighth note, instead of in hertz or milliseconds. It is calculated from the tempo, so an LFO's movement or a delay's echoes stay in time with the music when the tempo changes. At 100 BPM an eighth note is 300 milliseconds.

The LFO : Sync and retrigger Also in Delay

The threshold is the level above which a compressor starts turning the signal down. Below it, the transfer curve lies on the diagonal, where output equals input.

Compression : Threshold and ratio

Through-zero flanging sweeps the delay down to zero and past it, as tape flanging could by bringing the delayed machine level with the other. At zero delay the two copies are identical and every notch is above the top of hearing, so the sound is momentarily unfiltered. Some flangers imitate this by delaying the original slightly as well.

Flanger : Feedback and through-zero

A tie holds a sequencer's gate high across one step and the next, so the note continues without being restarted.

Gates and triggers : Gate length Also in Step sequencers

Time-stretching changes the length of a sound without changing its pitch. A granular instrument does it by moving the position the grains are taken from slower or faster than real time while each grain plays at its original pitch.

Granular : Position and pitch

On a delay, the tone control is a lowpass filter inside the feedback loop, so each repeat has a little less treble than the one before. It imitates the high-frequency loss of each pass through a tape machine or an analogue delay chip.

Delay : Feedback and tone

Tracking is how accurately an oscillator follows 1V/oct across its range. An oscillator that tracks accurately gives exactly one octave per volt, and one that tracks inaccurately plays in tune at the note it was tuned to and goes progressively sharp or flat further away. Manufacturers usually state tracking as an error in cents over a number of octaves.

One volt per octave : Tracking and calibration

A transfer curve plots a process's output level against its input level. A compressor's curve follows the diagonal below the threshold and bends to a shallower slope above it. A distortion's curve maps each input level to an output level.

Compression : Threshold and ratio Also in Distortion

Tremolo is a regular change in level. On a synthesiser it is an LFO sent to the amplifier's level.

The LFO : An LFO and three destinations

The tresillo is a rhythm of three hits across eight steps, spaced 3-3-2, and is common in Latin American music. It is one of the traditional rhythms that the Euclidean procedure produces.

Trigger sequencers : Rhythms from an algorithm

A triangle wave contains only the odd harmonics, with the level of harmonic n at 1/n² of the fundamental, so the third harmonic is a ninth of the fundamental. The levels fall so steeply that a triangle sounds close to a sine.

Waveforms : What each waveform contains Also in Harmonics and overtones

A trigger is a short pulse, a few milliseconds long, that marks only the moment something starts. Parts of a synthesiser that act once per event, such as a drum voice or a sample and hold, respond to triggers, and a clock is usually a stream of them.

Gates and triggers Also in Clocks

A trigger sequencer stores only whether each step fires, in a grid with one row per sound. On each step it sends a trigger to every sound whose row is switched on there, and it is the usual way of programming drums. Sixteen steps make one bar of sixteenth notes in 4/4 time.

Trigger sequencers Also in Step sequencers

An oscillator's tuning control sets the pitch it plays when its pitch input is at zero or has nothing connected. Under 1V/oct it fixes where the scale starts, and every volt above that raises the pitch by an octave.

One volt per octave

Twelve-tone equal temperament divides the octave into twelve equal frequency ratios of the twelfth root of two, about 1.0595 each. It is the tuning used by most Western instruments and by synthesisers by default, and it is one of many ways of dividing the octave.

Pitch and frequency : Notes, semitones and cents Also in One volt per octave , Where pitch comes from

A unipolar signal runs from zero up to its peak. A unipolar LFO moves a setting in one direction only and back: on pitch it bends the note up and back down, so the note played is the lowest pitch heard. A unipolar signal can be made from a bipolar one by scaling it to half its size and adding an offset of the same amount.

The LFO : Bipolar and unipolar

A VCA, or voltage-controlled amplifier, passes a signal through at a level set by a second signal on its control input. Its output is the input multiplied by the control signal: zero on the control input silences the output, and full control passes the input unchanged. It multiplies audio and control signals in the same way.

The VCA Also in Envelopes , Gates and triggers

Vector synthesis mixes four sound sources with a single two-dimensional control, usually a joystick. A corner of the control gives one source alone and the centre gives all four equally. Sequential Circuits introduced it in the Prophet VS in 1986.

Vector Also in Wavetables

Velocity is the strength with which a note was played, stored with each note. In a piano roll it is usually shown as a separate lane of bars underneath the notes, and it is often sent to the voice's level or filter.

Piano rolls : What else a piano roll stores

Vibrato is a regular change in pitch. On a synthesiser it is an LFO sent to the oscillator's pitch, and an LFO at around 5 to 7 cycles per second moving the pitch by a fraction of a semitone either side of the note sounds like the vibrato of a singer or string player.

The LFO : Rate and depth Also in The classic voice , Control signals and audio signals

A voice is the set of parts that produces one note, such as an oscillator, a filter and an amplifier with the envelopes and LFO that control them. A monophonic synthesiser has one voice, and a polyphonic synthesiser has several.

The classic voice Also in Build it in seven steps , Hocketing

Voice allocation is the way the notes of one line are shared between several voices. The notes can go to the voices in turn, at random, or by pitch range, so that low notes go to one voice and high notes to another.

Hocketing : Ways to share the notes

Volt per octave is a pitch standard for control voltages in which each volt added to an oscillator's pitch input doubles its frequency, raising the pitch by one octave. A twelfth of a volt, about 83 millivolts, is one equal-tempered semitone. Eurorack and most analogue synthesisers with control voltage inputs use it.

One volt per octave Also in Pitch and frequency

A voltage-controlled switch sets its position from the level of a control signal. A slow rising ramp sweeps it through the positions in order, and a random signal makes it jump between them.

Switches : Sequential and voltage-controlled

Wave sequencing plays a list of recorded waveforms one after another, each with its own duration, pitch and level, with crossfades between them. Each of the four sources on the Korg Wavestation of 1990 can play one.

Vector : The Prophet VS and the Wavestation

A wavefolder reflects any part of a waveform that goes past a threshold back towards the centre, as if the waveform had been folded over at that level. Raising the level of a sine or triangle going into it folds each cycle more times, which adds harmonics.

Wavefolding

A waveform is the repeating pattern an oscillator produces. Any waveform can be built by adding sine waves at the fundamental and at whole-number multiples of it, and the waveform's sound depends on which of those multiples are present and how quickly they get quieter going up.

Waveforms Also in What a signal is , Harmonics and overtones

The waveshape is the shape of one cycle of a signal. Two signals with the same amplitude and frequency sound different if their shapes differ, because the shape determines which harmonics the signal contains.

What a signal is : Waveshape Also in Waveforms

Waveshaping maps each input level of a signal to an output level by a curve. Distortion and wavefolding are both forms of waveshaping.

Distortion Also in Wavefolding

A wavetable is a set of single-cycle waveforms, called frames, stored in order so that neighbouring frames are similar. A wavetable oscillator plays one frame at a time, and moving its position through the set changes the timbre at the oscillator without a filter. The German instrument designer Wolfgang Palm developed the technique in the late 1970s, and his company PPG released the PPG Wave in 1981.

Wavetables Also in Vector

The position control of a wavetable oscillator sets which frame plays, or which two neighbouring frames are blended. Sweeping it with an LFO or an envelope changes the timbre continuously.

Wavetables : Frames and position

West coast synthesis is an informal name for adding harmonics to a waveform with few of them, such as a sine or triangle, by wavefolding and other waveshaping, associated with Don Buchla, who built synthesisers in California from the 1960s. It is contrasted with east coast synthesis, the subtractive approach of Robert Moog's instruments, and the two are often combined.

Wavefolding : West coast

White noise has the same power at every frequency, so its spectrum is flat when plotted against hertz. Each octave contains twice as many hertz as the one below it, so most of its power is in the treble and it sounds like a bright hiss.

Noise : White, pink and brown

XOR, or exclusive or, is a logic rule whose output is high while exactly one input is high. Applied to two rhythms it keeps the hits that belong to only one pattern and removes the ones they share, which tends to move the result off the beat.

Logic : Combining rhythms