L3 · Craft
Arrangement, sound-design depth, modulation, audio-reactive craft.
1,335 atoms · grouped by primary domain
A · Music theory & musicianship — 122
19-TET adds pitches between the cracks of standard 12-note tuning while keeping octaves aligned
A 3- or 5-step hat loop against a 16-step pattern creates an evolving polyrhythmic feel
A catalog of attributes lets you take inspiration from music without copying it
A displaced backbeat anchors the listener's meter, making other on-grid elements sound early or late
A dissonance curve plots sensory dissonance vs. interval for a given spectrum, with minima at consonant intervals
A dissonance score is a time-series graph showing how sensory dissonance ebbs and flows throughout a musical performance
A Euclidean rhythm is aksak (built only from duration-2 and duration-3 cells) exactly when 2k < n < 3k
A groove template is a timing and velocity map applied to shift any pattern toward a reference feel
A Hardstyle buildup has three phases: tease, melody preview, and tension-building filter sweep into the drop
A single-note, heavily-effected synth in the upper register raises energy without competing with the melodic content
A spectrum and a scale are 'related' when the spectrum's dissonance curve has minima at the scale's steps
A tuned kick layer with automated pitch can act as melodic percussion, not just a beat
Active listening — focusing on one parameter at a time — extracts usable technique from music
Alternating clave variants within a loop lengthens the cycle and reduces monotony
Alternating higher- and lower-pitched kick drums across a pattern is a hallmark minimal-techno groove technique
An avoidance list forces new creative territory by pre-committing to not use familiar techniques
Applying rhythmic tools to wrong materials or at wrong settings produces useful unexpected results
Arbitrary constraints reduce decision paralysis by eliminating valid options before work begins
Assigning each element a foreground, middleground, or background role creates three-dimensional musical depth
Bresenham's algorithm for drawing digital straight lines is an implementation of the Euclidean algorithm
Chaining single-change mutations creates descendants that diverge progressively from an ancestor idea
Chord-progression identification by ear trains hearing each chord's function as a harmonic sequence unfolds
Creating many one-change variants of a seed idea generates a sibling pool of musically related material
Creative resistance disguises distraction as legitimate work
Deep tool familiarity can trap you in the music your tools suggest rather than the music you want
Deliberately repeated microtiming deviations have greater groove impact than organic per-bar variation
Dilla time layers multiple simultaneous rhythmic feels — straight, swung, and displaced — within one track
Divisive rhythm subdivides a fixed cycle; additive rhythm builds sequences by adding or removing beats
Dragging (behind the beat) creates a heavy laid-back feel; rushing (ahead) creates urgency and forward drive
Dropping the bass one bar before a structural change signals the transition and eases the landing
Each genre has characteristic groove parameters: drag/rush amounts, swing ranges, and which elements vary
Effective counterpoint requires rhythmically independent voices that each function as standalone melodies
Equal temperament divides the octave into 12 equal semitones; just intonation uses pure frequency ratios — the two differ by cents
Euclidean rhythms are exactly the maximally even rhythms — those maximizing pairwise inter-onset distance on the circle of time
Exotic scales (Neapolitan, Middle Eastern, Hungarian, whole tone) extend the palette beyond major/minor
Exploring without a goal unlocks creative directions that task-oriented work closes off
Extended chords stack successive thirds above the seventh, and the thirteenth uses all seven scale degrees
Extracting a formal skeleton from an admired track provides an arrangement blueprint that avoids direct copying
Filling the entire arrangement with material first, then subtracting, avoids the blank-canvas problem
Finishing a track means arranging its looping parts into sections and bouncing the multitrack down to a two-track mix
Finishing one element to release quality before adding the next builds the skill of completion
Gamelan metallophones have inharmonic spectra, and their scales (pelog and slendro) are related to those spectra
Group humanization plugins link multiple tracks so instruments respond to each other's timing, mimicking an ensemble
Harmonic rhythm — the rate of chord change — is independent of surface rhythm and shapes perceived energy
Hearing intervals in tonal context requires naming both the interval distance and each note's scale degree simultaneously
Historically, 'consonance' and 'dissonance' refer to five distinct concepts, not one unified notion
Hocket distributes a single melodic line across two or more instruments to create timbral variety and rhythmic texture
Hood performed Minimal Nation by punching everything in live on a pocket sequencer for human feel
Hood structured Minimal Nation as a coherent narrative where each track is a chapter, not an isolated piece
Hood's production philosophy is 'pace yourself like a turtle' — momentum through endurance, not speed
House uses long DJ-friendly phrase forms: filtered intros, additive builds, chord/vocal breakdowns, kick+bass drops
Humanization addresses three independent axes: timing, velocity, and note duration — swing alone is insufficient
In techno a lowpass filter opening over 32 bars is itself the arrangement, replacing chord changes
Inharmonic timbres have a pseudo-octave — a non-2:1 interval that plays the functional role of the octave
Intertextual composition evokes familiar musical styles without citing sources explicitly
Introducing all elements early then removing them can create tension more effectively than holding them back
Introducing one micro-variation per bar across a drum loop prevents it from sounding like wallpaper
Jazz builds beats from the top (cymbals) down while rock builds from the bottom (kick and snare) up
Key modulation is a section-level event; treat a key change as structural, not a bar-to-bar move
Klangfarbenmelodie creates timbral variety in slow melodies by crossfading the same melody across different instruments
Layering different swing and quantization settings per track creates complex, organic groove
Layering material at half and double the track tempo creates rhythmic complexity within a single key
Linear drumming — no two drums play simultaneously — creates melodic-sounding beats and forces democratic instrument weighting
Listening in acoustically imperfect conditions reveals phantom elements and compositional opportunities
Listening to disliked genres reveals usable techniques and production approaches
Matching vocal rhythm to the natural stress pattern of lyrics produces natural-sounding vocal melodies
Meantone and well temperaments trade unlimited modulation for purer intervals and distinct key colors
Melody on chord tones sounds resolved; on tensions it pulls — good lines put tension on weak beats and resolution on strong beats
Minimal techno layers 'rhythms inside rhythms' so sustained listening reveals hidden structure
Minimising voice movement between chords through inversions creates smoother, more playable progressions
Minor ninth chords (m9) add a major ninth above a minor-seventh chord, producing a beautiful mellow sound used in deep house and ambient
Mirroring the intro, looping with subtraction, or fading out are three distinct strategies for ending a track
Modulation changes the tonal center mid-song; the dominant seventh of the new key is the primary agent
Non-12-tet equal temperaments (7-tet, 10-tet, 19-tet, etc.) are viable musical systems matched to specific timbres
Nu-disco arrangement uses drawn-out repetitive sections that ramp up and back, with filters and subtle changes keeping motion
Nudging the kick early or snare late creates urgency or drag without changing grid position
One borrowed chord from a parallel mode in an otherwise diatonic loop is a cheap, strong hook
One-time events — sounds, gestures, or processing applied exactly once — prevent loop-based music from feeling predictable
Optimal workflow is personal and must be synthesised through trial rather than copied from others
Parallel harmony transposes a fixed voicing by the same interval for each chord, producing the signature sound of house and electronic music
Partial quantization, selective instrument quantization, and manual velocity editing create human feel in programmed drums
Pelog is a seven-note Indonesian scale whose pitches do not correspond to any notes in 12-TET
Pink (1/f) noise has equal energy per octave and models the long-range correlations found in many natural signals including music
Postminimalism builds soundscapes through expressive timbre rather than harmonic complexity
Preparing the studio environment before inspiration arrives prevents creativity-killing interruptions
Pushing drums ahead of the beat drives energy; pulling behind relaxes the feel — set with track/channel delay
Quantizing kick and snare hard while leaving hats loose gives metric stability with textural feel
Randomisation tools require active curation — the producer remains responsible for every moment of the output
Recomposition discards most of an existing work then phases and loops the retained fragments into a new piece
Removing elements rather than adding them is often the path to fullness and clarity in a mix
Repeating a progressively shorter sub-loop of closing material creates a sense of acceleration that drives a track to a convincing end
Restraint — deliberately removing elements rather than adding them — is a compositional strategy, not a limitation
Reusing drum pattern MIDI as pitched material and vice versa creates rhythmically related harmonic and melodic parts
Separating creation from editing prevents premature judgment from killing creative flow
Sgubhu, a gqom variant, differentiates itself from standard gqom by using a consistent four-on-the-floor kick
Short looped fragments of any audio reveal implied rhythms through pattern-finding perception
Silence and noise are compositional materials as fundamental as pitched sounds
Simultaneously triggering an instrument with loops of different lengths creates a self-evolving composite pattern
Sketching many parts broadly before refining any one of them preserves fragile idea-generation momentum
Sleep music is an ultra-long, low-arousal ambient form built to accompany a full night's rest
Smooth voice leading keeps each chord voice moving by the smallest interval possible between successive chords
Staggering clip boundaries across tracks softens section transitions and avoids abrupt formal cuts
Starting an arrangement from a maximally dense section and subtracting layers is faster and more musical than building left to right
Starting from the bottom, from memory, or from an instrument breaks blank-slate paralysis
Step-function automation envelopes with sharp edges create an additional rhythmic layer independent of note events
Sufficient repetition makes almost any pattern feel musical regardless of its harmonic content
SuperCollider can generate a complete 12-tone row matrix using nested iteration over a scrambled pitch-class array
Swinging a bassline or melodic layer against straight drums generates groove without touching the drums
Techno's tension arcs span dozens of bars, driven by filter opening and layer subtraction rather than drops
Thai classical music uses approximately 7-tet because Thai instruments have bar-like spectra whose dissonance curves have minima near 7-tet steps
The complement of a Euclidean rhythm is also Euclidean
The consonance of an interval depends on the timbre of the sound, not just the frequency ratio
The dramatic arc — exposition, rising action, climax, falling action, dénouement — structures musical time at any scale
The hi-hat can function as an atmosphere-and-energy lifter, not only a timekeeper
The J Dilla drunk snare places the snare deliberately behind the backbeat
The leap-year patterns of the Jewish and Islamic calendars are Euclidean necklaces
Timeboxing breaks procrastination by making creative work feel manageable and stopping while it is still good
Transposition, inversion, retrograde, and pitch rotation are rule-based transformations that generate new patterns from existing material
Treating abandoned projects as a personal sample library removes the stigma of unfinished work
Treating kick and bass as a single monophonic composite line improves low-end clarity compositionally
Tuning electronic drums (especially kicks and toms) to the key of a song produces a more cohesive mix
Using simple or acoustic sounds while writing forces ideas strong enough to stand without production
B · Sound design & synthesis — 197
A dub techno kick is a 909-style sample with the filter lowered and release shortened to take its aggression off
A filter's frequency response describes its amplitude effect per frequency; its phase response describes the per-frequency time delay it introduces
A first-order Ambisonic B-format signal carries full-sphere spatial information in four channels: W, X, Y, Z
A good minimal record demands more production effort than a complex one because the sequence must improve with every listen
A Grain Delay pitched down an octave turns a chord's delays into the track's sub bass
A granular instrument requires the same daily practice as any acoustic instrument to sound compelling
A granular texture sounds the same played backwards because the grain is time-invariant
A granular voice reads a short window from a source table, applies an amplitude envelope, and outputs one grain
A harmonizer shifts pitch in real time by varying playback rate with spliced grain boundaries
A high-frequency ping-pong delay kept near-silent can be lifted in as a shimmering textural build
A highpass filter rather than lowpass keeps a synth chord thin and airy above the sub-bass
A long-decaying noise layer simulates room reverb within a clap patch without an external reverb unit
A noise oscillator layer gives delays textural material to sustain between note events
A periodic grain envelope creates AM sidebands at 1/grain-period intervals
A resonant filter's circuit topology, not just its cutoff, shapes its sonic character
A second carrier oscillator sharing the same modulator adds a formant region at any spectral position
A single modulated delay line produces chorus, flanger, and vibrato depending on its delay-time settings
A slow filter envelope attack creates a gradual harmonic swell that is key to the dub chord feel
A software emulation reduces an instrument's many possible sounds to a small labelled subset
A subtle 'colour' layer of modulation FX (chorus/tremolo/hiss) adds 3D dimensionality to dub — pick one or two per sound
A trainlet is a brief harmonic impulse train particle
A vocoder imposes the spectral envelope of a modulator signal onto a carrier using a bank of matched band filters
A wavetable oscillator must reduce its harmonic count as pitch rises
Ableton's Echo is the closest built-in device to a hardware dub delay; slight L/R timing error adds character
Adding a small constant to the FM modulating frequency creates a beat or tremulant effect
Additive synthesis builds a timbre by summing a fundamental with individually controllable harmonics
Alias-free wavetable oscillators use a set of per-octave band-limited tables that drop harmonics as pitch rises
An allpass filter leaves the magnitude spectrum flat but delays different frequencies by different amounts (phase dispersion)
An inverted modulator envelope creates a metallic release sound used in FM harpsichord and bell patches
An undamped bass-heavy soundsystem in a small room can hit the room's resonant frequency, causing mechanical feedback through objects on the DJ table
Applying a per-oscillator LPF or HPF warp lets two oscillators share frequency space without clashing
Applying waveshaping distortion inside the oscillator warp stage creates harmonically richer, fatter timbres
At high FM index or low c/m ratio, the modulated carrier's instantaneous frequency can go negative
ATK decoders output channels in counter-clockwise order starting from front-center; call .directions to verify routing
Audio round-trip latency has multiple additive sources and can be quantified with a loopback pulse test
Backing off a sampler's amp-envelope attack removes a drum sample's transient, turning it into a tonal element
Bass panned across the stereo field cannot be cut cleanly to vinyl and must be summed to mono
Binaural ATK decoders convolve B-format with HRTFs to simulate how the head filters spatial cues for headphone listening
Binaural audio uses Head-Related Transfer Functions (HRTFs) to simulate 3D spatial audio over headphones
Biquad filters are a two-pole two-zero IIR structure implementing most common EQ and filter types from six coefficients
Bit-crushing a sound then bandpass-filtering it turns harsh noise into usable dub texture
BufRd reads a Buffer at any frame index, enabling sinusoidal, noise-driven, and custom playback paths
Capping wavetable harmonics at the limit of hearing rather than Nyquist saves memory
Cascading identical biquad filters degrades the corner response — each stage needs a unique Q derived from pole spacing
Changing filter envelope slope shape (convex vs concave) via modulation matrix produces distinct filter sweep characters
Cloning one particle into a repeated stream builds a pitched tone from a single grain
Coloured noise is made by scaling each DFT bin of white noise by 1/f^B before inverse-transforming
Combine distortion, amp simulation, and saturation for dub 'colour' — but don't overcook, since mastering boosts it further
Concatenative synthesis drives resynthesis by selecting and joining database sound segments whose features best match a target
Coupling pitch bend and modulation wheels lets FM performers bend pitch while simultaneously darkening timbre
Critical distance is where direct sound energy equals reverberant sound energy in a room
Cross-synthesis imposes the spectral envelope of one sound onto another by multiplying their spectra
Cutting mid-frequencies in a Reese bass creates mix space for other elements like snares and leads
Defining two index breakpoints (I1, I2) and an envelope shape controls the full spectral trajectory of an FM note
Dennis Gabor proposed the grain as an indivisible psychoacoustic quantum of sound
Digital oscillators band-limit sharp waveform discontinuities to prevent aliasing
Distortion plus a high-feedback delay turns a dry vocal chant into an EBM industrial texture
Dub techno bass is deliberately simple — often a single sustained sine note pushed forward in the mix
Dub techno drums are low-passed to cut the top end and then saturated — the opposite of clean, punchy minimal-techno drums
Dub techno is a vibe and aesthetic, not a specific toolset — any synth becomes dub via saturation, a coloured filter, then delay and reverb
Dub techno melody ranges from old-school one-note to structured lines, usually in a minor key with D a common root
Dub techno sub bass is made from the kick sample by cutting its transient and keeping only the subby tail
Dub techno's live feel comes from slowly automating filter, delay feedback, and reverb decay across the section
Each grain can carry an independent spatial position for three-dimensional microsound projection
Each polyphonic grain in Max/MSP needs a unique instance ID to avoid shared-state collisions
Each Surge XT LFO exposes three independent outputs: full LFO, raw waveform only, and envelope only
Effects such as delays can be constituent parts of sound creation, not just post-processing
Electronic tracks stay engaging by having a sense of direction and propulsion rather than a static repeating loop
Emulate the dub mixing board by driving the delay/reverb chain into saturation — overdrive placed after the wet effects
Enabling pitch tracking on a noise oscillator makes it follow the played MIDI note for melodic noise effects
Extended techniques on acoustic instruments produce hybrid timbres unavailable through synthesis
Filters not only attenuate but phase-shift a waveform's individual harmonics, distorting its shape
FM and PM produce identical output at high sample rates; PM is preferred for digital implementation
FM brass timbres use c/m = 1/1 with index tracking amplitude, capturing the loudness-to-brightness coupling of brass instruments
FM percussive sounds use inharmonic c/m ratios with index decaying from dense to sparse spectrum
FM sideband amplitudes are set by Bessel functions of the first kind indexed by sideband order and modulation index
FM sidebands that fall below 0 Hz reflect into the positive spectrum with a phase inversion
FM woodwind timbres use a higher carrier harmonic and inverse index-amplitude coupling, causing higher harmonics to lead the attack
FOA transformations warp the whole B-format soundfield and are not commutative
FoaPanB encodes a mono signal to FOA B-format with a dynamically modulatable azimuth and elevation
FOF synthesis generates formant spectra from streams of grain bursts
Framing sound design in signal-flow terms makes it transfer across synths regardless of a synth's UI
Frequency shifting adds a fixed Hz offset to every partial, breaking harmonic ratios into inharmonic spectra
Glisson synthesis gives each grain an independent frequency trajectory (chirp)
Grain-by-grain pitch shifting produces chorus, noise, or melodic elaboration depending on grain size
GrainBuf granulates a Buffer with independently controllable trigger rate, grain duration, pitch, and position
Grainlet synthesis links any synthesis parameter to any other parameter
Grains shorter than about 50 ms transition from perceived pitch to timbre or click as duration decreases
Granular studio composition generates clouds with a synthesis tool then arranges them on a DAW timeline
Granular time-pitch changing in SuperCollider decouples playback speed from pitch by varying grain rate and position
Granulation enables independent control of pitch and duration
Grouping harmonically related partials under a single envelope reduces the parameter count in additive synthesis
Higher-order granulation re-granulates already-granulated sound to spawn new mesostructures
Holding FM modulation depth steady and enveloping separate spectrum bands gives more natural timbres than sweeping the index
Humans localise sound using interaural time, level, and spectral cues from the pinnae
In dub techno the effects chain, not the source sounds, does the bulk of the creative work
In Hardstyle, the kick drum must be pitched to match the track's melody — the most crucial production requirement
In Surge XT you can modulate one LFO's parameters with any other modulation source
Karplus-Strong physical modeling feeds an excitation signal into a tuned feedback delay to synthesize plucked or bowed strings
Layering detuned oscillators models the multiple resonant modes of a drum membrane
Leaving 2–3 dB of headroom or using true-peak metering prevents inter-sample overs in distribution
Mala's production philosophy was that contributing by removing elements is as valid as adding them
Microfiltration applies a unique filter to each grain for spectrally animated textures
Micromontage assembles micro-sound particles manually or algorithmically on a timeline
Microsound composition operates simultaneously on multiple time scales
Mixing on an analog console trains ear-based instinct that digital screens cannot replicate
Modulating partial frequencies and amplitudes with noise spreads each sine into a band, letting additive synthesis approximate noisy timbres
Morphing between two parameter sets gives additive synthesis a single performable control dimension
MTS-ESP allows one central source instrument to retune all compatible plugins simultaneously in a session
Multichannel reverb mix modulated by source-to-speaker distance models sound spatialization in SuperCollider
MUSIC V's unit generator model chains composable processing blocks to build synthesis instruments
Narrow EQ boosts on a synthesized drum can model the resonant chambers within the instrument
Only three frequencies — sr/3, sr/4, sr/6 — produce sine waves with zero quantization error at any amplitude
Overlap-add processing reconstructs audio from overlapping FFT windows without boundary artifacts
Per-grain processing applies a different transformation to each grain, producing granular heterogeneity
Perceptual quality depends on the group character of spectral evolution, not the precise amplitude curve of each partial
Physical model pitch is inversely proportional to delay time, requiring reciprocal computation from note number
Physical modeling synthesis approximates acoustic instrument physics using mathematical models of wave propagation
Pitch-synchronous granular synthesis aligns grain envelopes with the waveform period to reduce audio artifacts
Polyphonic MIDI in SuperCollider uses an array of 128 Synths indexed by note number
Polyphonic overlapping grains form a grain cloud — granular synthesis's characteristic dense texture
Procedural sound synthesis layers independent filtered noise generators to build one complex texture
Pulsar synthesis independently controls fundamental frequency and formant frequency
Quasi-synchronous granular synthesis uses near-equal grain timing to produce amplitude modulation
Randomising grain parameters between bounded min and max values produces animated granular textures
Recording unexpected outputs from unfamiliar processes creates source material that pure composition would not produce
Recording vinyl at 45 rpm into the SP-1200 and pitching down creates characteristic lo-fi grit that defines filter house texture
Resynthesis analyses a real instrument recording into partial tracks and uses those measurements to set additive synthesis parameters
Reverb is roughly half of dub techno's sound — heavy reverb that is modulated, filtered, and distorted, paired with delay
Rolling off the lowest bass frequencies in mastering can paradoxically make a vinyl record sound heavier
Running a complex oscillator into a VCA produces abstract metallic hat sounds beyond noise-source patches
Setting a DX7 operator to 1 Hz creates a sub-audio carrier that produces cyclic amplitude beating
Setting filter resonance at the edge of self-oscillation creates a dynamically expressive patch
Shaper applies a waveshaping transfer function stored in a wavetable buffer to an audio signal
Sidechaining the reverb send to the lead ducks the reverb during melody notes, keeping clarity while preserving space in pauses
Signal.sineFill creates additive synthesis wavetables by specifying partials as an amplitude array
Simple FM extends into a family of richer techniques via feedback, multiple operators, and modulated parameters
Six FM operators let a voice split its spectrum into three independently-enveloped bands
Soft saturation limits a too-hot signal more musically than hard clipping and leaves it recoverable downstream
Sound design can be guided by felt, tactile, and emotional qualities alongside technical parameters
Sound motion is organized as simple, complex, and compound processes that evolve spectrum, space, or both
Sparse, particle-based audio produces more event-legible visualisations than dense continuous music
Splitting one bass MIDI into high, mid, and sub layers produces a fuller, more controllable low end
Staggering harmonic attack times in additive synthesis simulates a low-pass filter sweep
Standing waves form when a sound's wavelength matches a room dimension, creating fixed nodes and antinodes
Subtle automation or randomisation of envelope and pitch parameters emulates the organic variability of live instruments
Subtractive analogue synthesis cannot realistically replicate an acoustic guitar
Summing a signal with a delayed copy creates a comb filter: evenly-spaced spectral peaks and nulls set by the delay time
SuperCollider's GrainSin, GrainFM, and GrainBuf UGens generate individual grains at a specified density
SuperCollider's Osc UGen requires a buffer in wavetable format, not a plain Signal
SuperCollider's SynthDef and Pbind together create a patterned synthesizer: SynthDef defines the voice, Pbind sequences it
Surge XT supports full microtonal retuning via Scala SCL/KBM files with two options for how pitch modulation tracks the tuning
Surge XT's Alias oscillator deliberately generates aliasing artifacts as a creative digital noise source
Surge XT's comb filter delays the signal by a pitch-tuned amount to create resonant coloration and physical-model waveguides
Surge XT's Ensemble effect models BBD bucket-brigade delay lines for authentic analog chorus character
Surge XT's FM2 targets musical FM sounds; FM3 adds a fixed-frequency third modulator for inharmonic timbres
Surge XT's Modern oscillator blends aliasing-suppressed saw, pulse, and triangle via DPW algorithm
Surge XT's MSEG provides fully editable multi-segment envelopes with looping, node types, and curve controls
Surge XT's Nimbus effect ports Mutable Instruments Clouds granular texture processing into a plugin effect
Surge XT's Tape effect physically models analog tape saturation, bias, speed, and playhead geometry
Surge XT's Twist oscillator ports the Mutable Instruments Plaits macro oscillator with 16 selectable synthesis models
Surge XT's Vintage Ladder filter models the Moog ladder circuit via numerical integration of the differential equation
Surge XT's Window oscillator multiplies a wavetable by a window waveform, creating formant-like timbres
Tech house kicks use a transient-heavy punch layer plus a pitched sub layer on every beat for low-end weight
The ATK provides two toolsets: FOA (first-order, 4-channel) and HOA (higher-order, (n+1)² channels)
The ATK uses a right-hand coordinate system: azimuth 0°=front, positive=counter-clockwise, elevation 0°=horizon
The ATK workflow is encode → transform → decode, each stage operating on B-format
The auditory system resolves frequency within critical bands: components within one band interact, components in separate bands do not
The dub techno chord is a one-note trigger expanded to a minor triad over detuned oscillators, filtered low with a tight envelope
The grain envelope shape determines the spectral spread and character of a grain
The granular paradigm is a universal representation for sound: any signal decomposes into time-frequency grains
The imprecision of analog gear at its operating edge is a sound-design asset, not a defect
The N2 denominator of the FM frequency ratio determines which harmonic series members are absent
The odd/even warp selects only odd or only even harmonics from a waveform, converting a saw into a square
The phase vocoder converts audio to a time-varying spectrum enabling pitch-time manipulation
The phase vocoder enables independent time-stretching and pitch-shifting by operating on FFT analysis frames
The simplest AR envelope is a gate fed through a one-pole filter whose coefficient sets attack and release time
The slope of the grain envelope controls the spectrum: sharper attacks produce broader bandwidths
The TR-808 hi-hat circuit mixes six oscillators through a bandpass filter and is impractical to replicate in modular
The TR-808's Manual Play mode schedules pattern transitions and fill-ins to happen at phrase boundaries, not immediately
The TR-808's trigger output can be routed as an audio signal to create additional percussive textures
Use envelopes for signal-reactive modulation and LFOs for constant movement — in dub, almost nothing stays static
Using a non-sinusoidal waveform for the highest partial fills in upper harmonics when oscillator count is limited
Using a sound only once, rather than looping it, makes a track a sequence of one-off surprises
Voice synthesis models the vocal tract as a series of formant resonances shaped by the source-filter model
VOsc morphs between two or more wavetables in adjacent buffers using a floating-point buffer index
VOSIM synthesizes vowel formants using decaying squared-sine pulse trains
Waveguide synthesis models acoustic resonators as digital delay lines with filters, efficiently simulating instrument resonances
Wavelet analysis uses variable-length windows for better time resolution at high frequencies
Wavesets segment a signal at zero crossings and enable a catalog of micro-distortions
Waveshaping synthesis passes a signal through a nonlinear shaping function, and input amplitude controls spectral brightness
Wavetable lengths are rounded up to a power of two for FFT and binary-index efficiency
Weightless strips grime's aggressiveness to create an atmospheric minimal subgenre that contrasts mainline grime's dense energy
Welsh's Synthesizer Cookbook teaches a 102-patch recipe method for systematic voice design on analog synthesizers
Wrapping a Pbind in Pdef allows live re-evaluation of patterns without stopping the stream
Xenakis's screens theory distributes grains on time-frequency planes using stochastic processes
XFade2 enables smooth blend between a dry signal and a bandpass-filtered wet signal in a SynthDef
C · Sampling, field recording & sample culture — 43
A double Mid-Sides array captures height information a stereo pair cannot
Alternating between near-identical slices of a repeated hit restores a natural, human feel
An off-beat crash cymbal falling between beats 3 and 4 marks the end of the Amen phrase
Appropriation is legitimate when the borrower 'betters' the source — Milton's criterion for creative transformation
Choosing source material is half the creative work of sample-based production
Close omnidirectional microphone placement outperforms distant directional microphones for wildlife field recording
Copyright terms extending to life-plus-50 years lock cultural material away from reuse for decades after an author's death
Culture jamming redirects corporate imagery back against itself to produce critical commentary using the original's own aesthetic codes
DAW groove extraction captures a live break's exact timing and velocity offsets and transfers them to a MIDI clip
Detuning and saturating a sampler kick helps it blend with a sampled breakbeat
Digital recordings fall off a cliff from presence into absence while analog recordings fade gracefully
Dillinja's DnB bass design treated the sub-bass entry not as a melodic line but as a 'one-note detonation' of stacked low-end timbres
Disabling time-stretching links pitch transposition to tempo change, creating rhythmic variation from a single sample
DnB producers switch between two different breaks each bar to create rhythmic variety and tension
Duplicating a snare slice and tuning each copy up produces the classic drum-and-bass ascending pitch fill
Early sampling culture predated rights-clearance, leaving source performers uncompensated
Extreme cold degrades cables and microphone capsules before electronic circuits fail
Field-recording performance works as a 'sound polaroid' and 'invisible map' that snaps listeners into their environment
Freesound's API supports content-based search filtering by automatically extracted audio descriptors
Hip-hop and electronic sampling proliferated during a brief window when copyright enforcement lagged behind the technology
In freeze mode, the playback position of a sample becomes a continuous sound parameter instead of advancing automatically
In pop music, timbre and production texture have replaced melody as the primary copyrightable identity
Jam sync timecode locks location audio to camera via a brief cable handshake and periodic re-checks
Journalists routinely reprint and cannibalize what others have written without checking sources — making media manipulation easy
Jungle tracks circulated primarily via acetate dubplates that wore out after ~50 plays, creating a hyper-local release cycle
Keeping a sample's surface noise is an aesthetic choice that anchors a track to an era and feeling
Lavalier microphones enable recordings from inside cavities and at extreme environmental conditions
Leaving multiple hits inside one slice preserves the original loop's groove better than slicing every hit
Long cable runs separate the recordist from the microphone, reducing mechanical noise and human disturbance
M-S decoding lets you adjust stereo width after recording by changing the mid-to-side ratio
Masking human scent on microphone windshields lets you place mics near scent-sensitive animals without disturbing them
Micro-power FM radio is the modern pamphlet — a community speech technology suppressed by spectrum-scarcity doctrine
Parody qualifies as fair use because it requires conjuring the original to comment on it
Plunderphonics proposes crediting source artists rather than seeking permission as the appropriate norm for transformative sampling
Recombinant plagiarism treats finished works as raw material, starting where others stopped without hiding the sources
Running digital recordings through analogue preamps adds a distinctive tonal character
Running redundant microphone paths protects against single-point failure in inaccessible recording locations
Speeding up a loop to resample it then returning to the original tempo adds authentic grit
Strict sampling law creates a two-tier system: artists rich enough to clear samples, and outlaws who can't afford to
Technostalgia is not nostalgia for the past but a strategy for achieving a particular present sound
Thru playback continues past a triggered slice into the rest of the sample for natural breakbeats
U2 performed the same real-time appropriation on their ZooTV tour that they sued Negativland for, exposing appropriation law as power-asymmetric rather than principled
Velocity shading of double-hit kicks and ghost snares shapes breakbeat feel more than note placement alone
D · Mixing, mastering & loudness — 88
A -7 dB cut at 600 Hz with a wide Q creates the hollow mid-scooped character of dub techno chords
A 1/3-octave graphic EQ provides enough resolution to notch room resonances without disturbing adjacent frequencies
A ducker gives even keyed rebalancing where a keyed compressor over-reduces loud notes
A gate or volume LFO triggered to tempo can create rhythmic gain pumping in sync with the groove
A parametric equalizer allows independent control of center frequency, gain, and bandwidth (Q) for each band
A polarity-inverted parallel gate channel can implement sidechain ducking without a dedicated ducker plugin
A rough mix establishes the balance direction the artist expects; deviating too far requires client approval
A Schroeder reverberator builds artificial reverb from parallel comb filters and series allpass filters, with mutually-prime delay times
A slow-attack field-recording layer swells behind the kick to add breathy organic texture without a transient
Adding a tiny manual track delay to one layer of a layered clap creates a looseness without full humanization
Alternate mixes and stems give producers and labels options without requiring a full recall
An EQ→saturate→EQ chain on a kick adds harmonic character while controlling the resulting aggression and transients
Analogue-modelling plug-ins degrade outside their intended operating range because they don't fully model extreme non-linearities
Arrangement clarity comes from creating space — fewer competing parts, not more layers
At high frequencies a fraction-of-an-inch position change flips a comb-filter peak into a null
Automating or modulating a filter on a delay return channel adds evolving movement to the wet signal
Automating section-to-section levels preserves the dynamic arc that makes a song feel alive rather than static
Bouncing tracks down on a limited tape machine permanently fuses them, so sound separation cannot be recovered later
Buss compression glues a mix and should be inserted during mixdown so its side effects shape mix decisions
Chaining compressors lets each stage handle a different dynamic task
Chaining two compressors in series allows independent control of different dynamic problems on the same signal
Combining one tempo-synced and one free-running delay tap creates polyrhythmic dub delay texture
Controlled distortion adds harmonics that make quiet or masked sounds more audible without changing their frequency
Correct overall mix tone with broad master-buss EQ, keeping narrow moves on channels
Delays often outperform reverbs in dense modern mixes because they occupy less space while achieving the same blend effects
Deliberately not mastering a track to maximum loudness preserves the dynamic range that lets bass hit physically
Detailed fader rides push weak syllables up to even out a vocal for intelligibility
Dialnorm metadata sets a consistent program loudness reference in Dolby Digital, defaulting to -27 and running -31 (loudest) to -1
Drum sound replacement doubles (not replaces) subpar drums with samples to improve consistency while preserving human feel
Dub techno percussion is pushed into dub territory with amp distortion, delay, and added noise
Dub techno's space comes from two shared sends: an overdriven reverb and a 100%-wet ping-pong echo
Editing non-overlapping repetitions of a single track to create a fake double-track adds width without phase cancellation
EQ boosts on multimiked tracks change comb-filtering relationships unpredictably; cuts are safer
EQing a parallel dynamics return aims the processing at specific frequency regions
Experienced mixers hear the finished product in their heads before touching a fader
Fader automation adds mix dynamics by riding individual elements — what a live band does naturally must be recreated deliberately in the DAW
Frequency masking between competing instruments is best resolved by ear with manual fader automation rather than by keyed processing alone
Garage drum beats need light-touch bus compression because the genre depends on dynamic range
Give each element its own ambient environment so reverbs and delays do not clash
Glue a drum kit with gentle bus compression, letting the sculpted sounds lead rather than heavy-handed processing
High-passing a reverb's input keeps low frequencies out of the tail and prevents low-end mud
Inter-sample clipping can occur when the reconstructed waveform exceeds 0 dBFS between sample points even if no individual sample clips
Keeping the stereo field deliberately narrow in a club mix helps low frequencies translate to mono sound systems
Layering the same chord through a different delay time adds cross-rhythmic depth without harmonic conflict
LCR panning places all tracks either hard left, center, or hard right to produce a clean, uncluttered stereo image
Loudness maximization is a chain of stages that each trade dynamics for loudness with characteristic artifacts
Master-buss compression applied early in the mix allows individual tracks to be balanced against the buss processing
Mid/Side encoding separates the mono centre from the stereo-only sides for independent control
Multi-band tape saturation on the drum bus shapes frequency balance and adds perceived loudness
Multi-stage compression across analog and digital domains achieves high levels with less audible artifact than a single heavy stage
Multiband compression is for imbalances specific to one frequency band
Multiband dynamics fix a dynamic problem in one frequency region without touching the others
Multiple distinct techniques for widening stereo use different mechanisms and have different mono-compatibility profiles
Parallel compression blends a heavily compressed copy under the dry signal to add sustain while preserving transients
Parallel distortion adds harmonic density to a specific frequency range without altering the dry signal's dynamics
Perceived loudness comes from arrangement space and saturation — a mix with gaps sounds louder than a full, crushed one
Pitch correction should fix obvious errors while preserving the micro-variations that make a voice sound human
Playing on a large sound system immediately exposes production problems that studio monitors mask
Process only when you can name what you are hearing that needs to change
Processing all drums together on a shared bus glues them into one cohesive instrument
Processing parameters that work in the chorus may need automation to avoid problems in verses and breakdowns
Purely tonal EQ (for subjective appeal rather than balance) can freely use boosts and analogue character
Rebalancing premixed loops/samples relies on editing, phase-cancellation and selective processing
Reverb and delay effects can be widened using MS or other stereo-enhancement techniques on the return channel
Rock arrangements balance guitar, bass, and drums co-equally, unlike bass-led hip-hop and pop
Sample-peak meters miss inter-sample peaks, which can exceed 0 dBFS by 3 dB and require an oversampling True-Peak meter
Saturating the drum bus low end boosts perceived bass and loudness more gently than an EQ boost
Sidechain-ducking mid-range instruments under the vocal clears space when the vocal is present
Spotify cannot lift a quiet master to -14 LUFS if True Peak headroom is insufficient
Spotify normalizes albums as a unit but normalizes individual tracks when shuffled or in playlists
Spotify normalizes all tracks to -14 dB LUFS at playback using the ITU 1770 standard
Spotify's True Peak ceiling for masters is -1 dBTP, tightening to -2 dBTP for masters hotter than -14 LUFS
Spreading gain across multiple converters and level controls allows maximum loudness with minimal distortion
Streaming loudness normalization adjusts gain at playback only — the uploaded file is never altered
Streaming loudness normalization weakens the payoff of extreme limiting
Summing bus overdrive occurs before the master fader — pulling master fader down does not fix it
Tape echo delay is a defining element of dub aesthetics across reggae and dub techno
Tape machines running at high speed roll off steeply below 50 Hz, shaping the spectral signature of 1970s–80s rock
The Abbey Road reverb EQ curve (roll off below 600 Hz and above 10 kHz) makes reverb blend smoothly without muddying the mix
The elliptical equalizer sums low-frequency energy toward mono to prevent groove damage during vinyl cutting
The Interest element — finding and emphasizing the song's most compelling focal point — separates great mixes from merely good ones
The mix engineer steers the listener's single-focus attention to what matters each moment
The playback system a producer designs for shapes the basslines they write
Tuning a sound system means jointly optimising acoustics, dynamics, crossover points and driver selection
Use as few multiband bands as the problem needs, and set crossovers deliberately
Varying dryness across vocal layers creates front-to-back mix depth, not just width
Vinyl mastering must account for the cartridge compliance of the intended playback system
Volume automation directs listener attention and polishes balance problems that static processing cannot solve
E · Modular, Eurorack, grooveboxes & dawless — 121
A biodata module turns any biological signal into Eurorack CV, gates, and audio, making living organisms into performers
A Buchla sequencer jump can carry a certainty percentage, executing the branch only some of the time
A capacitance proximity detector turns hand distance into control voltage for touchless gestural control
A DAWless live techno rig distils a large studio down to a compact standalone hardware subset
A hacked game controller provides a cheap USB interface for connecting custom sensors to music software
A highpass filter after FM oscillators removes the tonal fundamental, converting FM output into metallic texture
A live modular rig can be built around spontaneous algorithmic sequencing or around pre-programmed banked sequences
A live techno rig should keep producing sound with no input so the performer manipulates rather than rebuilds
A live techno rig stays minimal because fewer voices mean less to manage while improvising
A lowpass filter can double as a mixer channel by treating cutoff-open as full send and cutoff-zero as mute
A Maths channel self-cycling at audio rate is an oscillator, and feeding another oscillator into its EOC output jack alters its tone differently than into Signal IN
A matrix mixer with output-to-input feedback self-oscillates into an instrument with no source
A mechanical plate reverb driven by solenoid percussion creates a hybrid acoustic-electronic instrument
A multi-output knob module used as a macro controller lets one gesture simultaneously animate multiple unrelated parameters
A piezo disk driven through a step-up transformer resonates physical objects, turning them into sculptural reverb units
A precision adder sums CVs at exact voltages to transpose a sequence by musical intervals
A quantizer applied to a smooth CV source creates machine-like stepped modulation
A sample-and-hold random voltage source injects controlled randomness into pitch, amplitude or timing
A single oscillator routed through waveshaper, notch filter, and granulizer can produce an unlimited range of sounds in a live modular rig
A techno rumble kick is the kick's own reverb tail filtered to sub-bass and re-shaped by an envelope
A VCV Rack cable can carry up to 16 voices; a mono modulation source fans out to all voices
Audible generative processes should change gradually so listeners can track the transformation
Buchla's 1966 system included light and projector modules to control visuals from the same patch
Buchla's Multiple Arbitrary Function Generator acts as one module doing many jobs, far beyond a sequencer
Capacitive touch sensors report finger position by centroid-weighted averaging across multiple activated pads
Control All applies a parameter change to all audio tracks simultaneously
DEJA VU is a probability knob that fades continuously between fresh randomness and a locked loop
Dub techno chord voicing stacks fixed semitone offsets to build a minor chord from one incoming note
Each Buchla 200e module type has a module ID that lets the preset manager and internal bus address multiple identical modules independently
Each track can have an independent length and speed to create polymetric patterns
Each VCV Rack cable adds a one-sample delay, which can cause clock/reset ordering bugs
Feeding a MATHS output back into its own CV input breaks the linear response, giving shapes Vari-Response cannot
Feeding an analog video signal into an audio amplifier makes a fixed-pitch drone whose overtones encode the image
Fill mode temporarily activates FILL-conditional trigs for a pattern variation
FM between two detuned oscillators produces inharmonic noise suitable for hi-hats and cymbals
Global Mute affects all patterns simultaneously while Pattern Mute affects only the active pattern
Hardware sequencers lack a shared protocol for jumping song sections together, so producers fake structure by muting tracks
Harmonically complex FM waveforms (saws, squares) generate richer inharmonic sidebands than sines for percussion
Heavy distortion on an analog audio signal creates a pseudo-square wave that CMOS digital circuits can process as a clock signal
In a live modular techno rig most modules serve sequencing and manipulation, not sound generation
In a live set reverb is played as an expressive element for builds and breakdowns, not a static effect
In VCV Rack 0 dBFS equals 10 V, and 0 dBVU sits at −18 dBFS by hardware convention
LFO modulation destinations span all sound-shaping parameters including other LFO parameters
LFO modules used as audio-rate partial generators extend oscillator count cheaply at the cost of aliasing and limited harmonic control
LFO trig mode controls whether the LFO restarts, holds, or runs free on each note
Logic gates (AND, OR, comparators) in Eurorack combine trigger and gate signals to create conditional rhythmic patterns
Low-pass filtering a reverb tail keeps only its bass, turning reverb decay into a sub-bass rumble
Marbles separates random rhythm (t) from random voltage (X) as two independently controllable stochastic layers
Marbles' Y output is a slow smooth random source, immune to DEJA VU, ideal for self-patched modulation
Mastering a small, specific tool set can produce deeper work than a large gear inventory
MATHS builds a full ADSR by chaining Channel 1 and Channel 4 with EOR as the link
Maths compares two signals and outputs a gate when one exceeds the other via SUM-based subtraction
Maths creates an Arcade Trill complex LFO by triggering CH.1 from CH.4 EOC and patching CH.4 output back to CH.1 Both IN
Maths divides an incoming clock by a ratio set by the RISE parameter of a triggered channel
MATHS divides an incoming clock by setting Rise time longer than the interval between triggers
MATHS extracts a gate from a CV by comparing it to a threshold and firing an instant EOR pulse
MATHS implements a 1-bit set-reset flip-flop with CH1 Trigger as Set and BOTH CV as Reset
Maths mirrors a control voltage around an offset point by inverting it on one channel and summing a fixed offset from another
Maths multiplies two control signals by patching both to a channel with RISE full CW and FALL full CCW
Maths performs full-wave rectification by multing a signal to CH.2 (normal) and CH.3 (inverted) into OR OUT
Maths simulates a bouncing-ball physics envelope by chaining two channels with decaying amplitude
Maths soft-syncs to a sawtooth oscillator by patching it to the lag input in audio-rate cycling mode
Micro timing shifts individual trigs ahead of or behind the beat grid
MIDI tracks use the same p-lock and condition system to sequence external gear with up to 4-note chords
Mixing a dry kick and a processed copy on separate channels lets each be balanced independently
Modulating the decay envelope length with a clock divider creates alternating open and closed hi-hat patterns
Multi-purpose voice modules keep a live rig compact by each covering several sonic roles
Nakamura treats the no-input mixer as an equal partner, surrendering control to what the system produces
Patching an oscillator to Maths trigger input and mixing EOR with the main VCO generates sub-harmonics
Patching Maths Variable OUT back to RISE or FALL CV independently controls each slope's response curve
Pattern chains sequence multiple patterns for one-shot playback without a saved song
Pattern trigger sequencers that run from a clock alone give an always-on rhythmic foundation
Performance modulation sources each route to up to four parameter destinations with independent depth
Poking contacts on an LCD display with a battery wire causes segments to distort and produces audible signals from the display's own oscillator
Portamento on the Digitakt II offers constant-rate and constant-time glide with legato and gating options
PRE and NEI conditions chain trig outcomes across steps and tracks
Preset locks swap a track's sound mid-pattern to a different preset from the pool
Raw modular output requires EQ, limiting, and compression to be mix-ready, just as any digital production chain would
Recording a live modular set with no overdubs or edits as an album creates a distinct purity and performative quality
Reliable trigger detection uses two thresholds (hysteresis) so one rising edge fires only once
Retrigs subdivide a single trig into a rapid burst of re-fires at a set rate
Routing a dry synth's audio through a sampler adds onboard effects without extra mixer channels
Running a metallic percussive output through a granular engine adds a sustained noisy texture to a minimal techno patch
Running an envelope through a resonant filter (instead of a slew) adds an LFO-like movement to the envelope's sustain and release phases
Running filter fades inside the performing rig removes the need to rely on the house mixer for energy control
Sample-and-hold and smooth random voltage sources animate patch parameters generatively so a static patch keeps evolving
Sidechaining voices to the kick ducks them on each hit, producing techno's pumping groove
Single-encoder interfaces trade immediacy for preset recall accuracy by eliminating pot-position mismatch
Song mode arranges patterns into a linear sequence with per-row tempo and mute settings
Splitting a bass into high and low bands and sending only the highs to a delay keeps the low end clean while adding space
SPREAD shapes the random-voltage probability distribution from a constant, through bell and uniform, to bimodal gates
STEPS carves notes out of a scale in order of least-used first, so one knob thins harmonic density
The 222e Kinesthetic Input Port tracks performer hand position in 3D space via IR rings, outputting XYZ control voltages for live gestural control
The 227e System Interface voltage-controls spatial placement of signals in 2D quad space and offers a swirl mode for rotating audio
The 230e envelope follower's transient pulse mode generates triggers only when a CV increase occurs — useful for extracting rhythmic info from complex audio
The 251e sequencer's per-stage loop counters enable nested rhythmic repetition without cycling the entire sequence
The 252e's ring display has five independent layers for pulses and CVs, enabling selective phase-shifting of rhythm, pitch, and velocity independently
The 256e CV processor applies piecewise-linear transformations to control voltages using a configurable breakpoint, enabling non-linear CV shaping
The 259e Twisted Waveform Generator's memory bank wavetables scan volatile program memory, producing unpredictable and non-repeating timbres
The 281e Quad Function Generator's quadrature mode chains two generators so their attack and decay phases overlap in a fixed sequence
The 285e frequency shifter shifts all harmonics by a fixed Hz amount, destroying harmonic relationships and creating inharmonic sidebands
The 291e Triple Morphing Filter interpolates between stored filter snapshots over time, with per-stage timing control
The 297 Infinite Phase Shifter creates barber-pole phasing — a comb filter whose notches continuously move in one direction without net positional change
The base-width filter defines a frequency band by center and bandwidth rather than cutoff and slope
The Buchla 252e Polyphonic Rhythm Generator sequences pulses and CVs on concentric rings driven by up to three independent clocks
The Buchla 260e generates Shepard-tone auditory illusions — pitches that appear to rise or fall endlessly without net movement
The Buchla 266e Source of Uncertainty generates multiple flavors of controlled randomness — fluctuating CVs, quantized random voltages, and stored random voltages
The Digitakt II Euclidean mode uses two independent pulse generators with Boolean logic
The Digitakt II's master compressor supports flexible sidechain routing including individual track sources
The MATHS OR output passes whichever channel is currently highest, acting as a maximum-voltage selector and half-wave rectifier
The Roland TR-808 cymbal uses six detuned square waves mixed and heavily filtered to create metallic noise
The Slice machine lets you manually define and sequence individual sample regions
The techno kick has a pitch that every other voice must be tuned to
Tides frequency-multiplication output mode makes just-intonation chords at audio rate and polyrhythmic triggers at LFO rate
Tides is calibrated for accurate pitch tracking by feeding known 1.000V and 3.000V references into V/OCT in a button-triggered routine
Tides uses different clock-following algorithms at LFO vs. audio rate: phase-locked for swing at LFO, pitch-accurate at audio
Trig conditions let each sequencer step fire only when a logical rule is true
Tuning analog drum modules live — pitch, decay, and saturation — is a primary performance gesture in modular techno
Turning a reverb's diffusion fully down converts it into a delay, usable as a dry-free parallel effect
Two Maths channels cross-triggered produce 90-degree phase-shifted LFOs (quadrature mode)
Werp, Stretch, and Repitch machines each use distinct algorithms to tempo-sync samples
F · Live coding: music — 241
A breakdown must drop energy to a genuine low so the following build has real contrast
A browser Sonic Pi can drive collaborative jamming by streaming server events to each client's Web Audio synth
A build must keep at least one voice in reserve to subtract at the drop
A drop is one cycle of near-silence followed immediately by the full groove returning
A fill applied only on the turnaround cycle is self-clearing and requires no follow-up save
A groove must be stated for a full phrase before it is developed
A high-pass sweep removes low-end weight from the full stack without removing the pattern
A highly constrained live coding language reduces cognitive load and enables musical output within seconds, at the cost of expressive range
A JITLib proxy is a placeholder defined by its valid sources, its contexts of use, and a default source
A live-coding setup file pre-loads server memory, sample dictionaries, SynthDefs, a clock and a safety limiter before the set
A live-coding SynthDef names its frequency arg 'freq', exposes an 'out' arg, and self-frees via doneAction
A livecoding agent must not claim to hear or see its output when no perception bridge exists
A low-pass cutoff ramped upward over a phrase opens the groove without adding voices
A pattern must be tagged .analyze('hydra') to be visible to the audio-visual bridge
A satisfying live set has an energy arc of contrast over time, not sustained maximum
A Strudel Hap has an onset when its whole timespan begins at the same time as its part timespan
A SuperCollider pattern is a factory for a stream, with Pseq and Prand giving ordered and random values
A wholesale section rewrite is only warranted when at a section boundary, the target shares few voices, and a patch would cost more edits
Adding a feedback delay on a voice's final cycle lets its echo tail carry the transition
Additive synthesis in Sonic Pi layers multiple synth voices at different pitches and amplitudes to create new timbres
Algorithmic composition creates rule systems that generate musical output rather than specifying individual notes directly
Algorithmic composition strategies range from pure randomness to Markov chains and constrained search
Algorithmic notation turns score-writing from description of intent into a process of live exploration
Algorithmic spatialization places sounds in virtual acoustic space using channel-based diffusion, object-based rendering (VBAP/Ambisonics/WFS), or binaural techniques
Algorithms extend compositional cognition by executing implications the composer cannot fully predict
An agent without the L3 perception bridge must not claim to have evaluated perceived audio or visual output
An autonomous live edit should carry exactly one concept-id and never bundle two ideas into one save
An edit is only sonically effective if its diff touched a sound-producing token, not just a comment or whitespace
Autonomous livecoding agents must apply stricter cadence discipline than copilot mode, defaulting one tier smaller
Autonomous livecoding mode uses a priority-ordered if-then table where the first matching row fires
Autonomous mode risks over-acting while copilot mode risks displacing the human performer
Breaking things and recovering fast is core livecoding practice, not a failure to hide
Bricolage programming is a creative feedback loop of making a code change, perceiving the output, and reacting — without forward planning
chop and striate slice a sample into n pieces; striate interleaves slices from multiple samples where chop plays each sample's slices sequentially
Chopping a sample into equal slices and playing them in random order creates beat-slicing effects
ChucK shreds are sample-synchronous concurrent processes spawned with spork ~
ChucK Unit Analyzers extract audio features from a signal for real-time analysis
Combining X-write with a movement operator animates a self-propelled operator across the grid
Commenting a voice out or in is the most legible and reversible layer transition
const replaces the entire playing pattern with another one
Continuous random patterns must be sampled by a discrete structure to produce events
Copilot mode follows a propose-explain-wait loop and never saves without human acceptance
CTK objects populate a Score-like structure that plays in both real-time and non-real-time synthesis
Curly brace syntax in Tidal creates polymeter patterns where sequences of different lengths share the same pulse
Designing a live-performance interface is instrument building — the mapping shapes what music is possible
Different live coding systems model time differently: metric cycles (Tidal), coroutines with explicit waits (ChucK), and temporal recursion (Extempore)
Driving percussion above 0 dB into StageLimiter compresses the whole mix on each hit for a sidechain-style pump
Each arc should reserve its single highest-energy move for one boundary it has built toward
Each livecoding edit should introduce or retire exactly one concept-id so the diff is legible and attributable
Estuary terminal commands add custom sample banks to an ensemble session
Estuary's Metre widget visualises cycle subdivisions and Euclidean/Bjorklund patterns against the live tempo
Every evaluation heuristic is either a static [NOW] check or a [L3] check blocked on a perception bridge that does not yet exist
fastcat and cat concatenate patterns in series, differing only in whether they compress to one cycle
filterHaps tests the whole Hap (with timing) while filterValues tests only the value
For time-based arts, the most useful definition of 'declarative' is closeness of mapping between notation and target domain
From-scratch live coding starts from an empty editor and builds the whole piece live before an audience
Generative composition means defining a rule-system and running it, not fixing every note
Glicol uses the LCS algorithm to update only changed nodes when code is re-evaluated
glicol_synth is a standalone graph-based Rust audio DSP library usable outside Glicol
Glicol's JavaScript API embeds the audio engine in any browser app via CDN or NPM
Glicol's meta node allows inline Rust DSP code inside a live session
Granular synthesis constructs sounds from thousands of short grains (10–100 ms) with independent parameters
High-level abstractions reduce cognitive load in live coding but may constrain musical affordances; some practitioners prefer lower-level control
Imperative engines can express tension arcs as continuous automation; cyclic engines can only produce section-quantized staircases
In ixi lang musical phrases are named agents you manipulate by name with verb-noun-adjective actions
In SuperCollider, effects buses require correct node ordering so the effect Synth reads the source Synth's output
In this livecoding rig a file save is the performance action: Strudel hot-swaps gaplessly and Hydra re-evals on the same GL context
In Tidal each sound parameter is patterned independently, so timbre aspects can carry their own polyrhythms
Interactive music systems react to performer input in real time, ranging from deterministic score-following to autonomous improvising agents
Interference patterns in live coding produce outcomes that exceed the coder's prior imagination
ixi lang encodes rhythm as monospaced text agents where horizontal position of characters marks timing
ixi lang operates as a 'parasite' on SuperCollider — adding a constrained high-level syntax that coexists with full SC access in the same document
ixi lang organises musical notation into three real-time-synchronised modes — melodic, percussive, and concrète — each with its own bracket syntax
ixi lang represents music as agents performing scores that rewrite themselves in real time when transformed by code
ixi lang rewrites an agent's score in the editor when it is transformed, keeping the code a live view of current state
ixi lang's 'coder' mode removes the latency between typing and sound by mapping keystrokes directly to audio in real time
JITLib can crossfade a proxy's source when it is redefined instead of cutting abruptly
JITLib treats a running SuperCollider program as an ongoing construction rather than a compile-then-run tool
Laptop music inherits acousmatic listening without solving the agency-perception problem that acousmatic music raised
Live coders find ensemble performance easier because performers can cover each other's gaps, reducing the solo exposure problem
Live coders work at least one level of abstraction above individual sound events, manipulating compositional structure rather than note-by-note detail
Live coding exists on a spectrum from fully improvised to largely pre-composed, and performers position themselves on it
Live coding intervenes in an ongoing process by modifying its laws (the program text), not its immediate state
Live coding makes software strange (defamiliarization), letting us see beyond routine practices and assumptions of computational culture
Live coding produces multiple overlapping kinds of liveness — machine liveness, embodied human liveness, and relational audience liveness — not one single kind
Live coding systems achieve progressive evaluation through external grounding, internal grounding, temporal recursion, or hybrid state management
Live coding's real-time loop between coder and machine is itself a creative method, not just a format
live_audio creates a persistent named audio input stream that can be moved between FX contexts dynamically
live_loops auto-generate cue events that other live_loops can sync on for phase alignment
Live-coding pattern languages are composable: any two transformations combine into a new one without extending the language
Livecoding edit cadence should be anchored to the musical grid, not wall-clock time
Livecoding edits span a four-tier diff-size ladder from micro param changes (1 line) to section rewrites (whole file)
Livecoding requires changes to happen live in the running file, not staged offline and dumped in as a batch
Locking a TidalCycles delay sets its time in cycles so echoes stay in phase with the tempo
Markov chains model context-dependent musical choices by making each event depend probabilistically on prior states
MIDIOut sends MIDI from SuperCollider to a DAW via the macOS IAC Driver virtual MIDI bus
MiniTidal's continuous signals (sine, saw, tri, square, rand) can modulate any control parameter
Naming a pattern-transformation function adds it to a system's compositional vocabulary and shapes what music is possible
Network latency above 20-30 ms disrupts tight rhythmic ensemble playing but can be worked around with free improvisation or asynchronous structures
Network music performance has latency thresholds: below 10ms is unnoticed; 10-40ms requires tempo negotiation; above 40ms demands independent tempi
NodeProxy and Ndef are SuperCollider's server-side placeholders whose synthesis source can be swapped while playing
NodeProxy.map connects a running control proxy to a synth argument without stopping synthesis
On a broken or silent output the agent must recover to a known-good state before adding new ideas
Only MIT/BSD/Apache/CC0/CC-BY non-NC code may be emitted verbatim into committed jam files; gray-licensed snippets are local-reference only
onset: pick in Sonic Pi selects a random transient event from a sample for instant hit variation
Orca's O and X operators read and write cells at an (x,y) offset from themselves
OSC lets SuperCollider send and receive timestamped musical messages over a network
Overlaying a structural pattern and a material (colour) pattern produces an interference result you cannot read off the code
Pbindef modifies individual keys of a running Pbind without restarting the stream
Pbjorklund2(k,n) emits Euclidean duration patterns that can be modulated by nesting other pattern classes
Pchain and Ppar compose event patterns by layering key sources or running multiple patterns in parallel
Pdef and Tdef are SuperCollider's client-side proxies for live-replacing running patterns and scheduled tasks
Pdefn holds a named pattern value that can be derived and reused across multiple running streams in SuperCollider
Performing live coding requires accepting public errors and resisting the pressure to meet conventional definitions of music
perlin produces smooth cycle-to-cycle random modulation unlike the per-event jitter of rand
Pfunc generates pattern values by calling a function, and Pkey cross-references values from an earlier key
Physical modelling synthesis uses an excitation signal driving a resonant body model
Placing with_fx inside a loop creates one FX instance per iteration, wasting CPU
play, synth, and sample return a SynthNode reference that enables post-trigger parameter changes
ply n repeats each event in a pattern n times within the event's original time slot
ply subdivides each event into n repeats; segment resamples the whole pattern at n events per cycle
polymeter aligns patterns by step so patterns of different lengths phase against each other
ProxySpace NodeProxies let you rewrite a running synthesis graph mid-performance with an automatic crossfade
ProxySpace turns the SuperCollider environment into a namespace of live-redefinable audio proxies
Pure algorithmic generation tends toward uniformity unless counteracted by entropy variation, interactivity, or inherent structure
pure() lifts a value into a pattern that emits it once per cycle
Quarks extend SC's language with community-contributed classes; UGen plugins extend the audio server with new signal processors
Raising degradeBy over successive cycles thins a voice while preserving its timing grid
randcat plays patterns from a list in random order; wrandcat adds probability weights to the selection
randrun emits a shuffled run of the integers below n, one distinct value per step
Redefining a named function while a named thread loops is the foundational Sonic Pi live-coding pattern
Renardo defines audio effects as SC SynthDefs in Python using an `effect_manager` DSL, then exposes them as player attributes
Renardo supports just intonation as an alternative tuning system via `Tuning.just` on any scale
Renardo supports microtonal pitches as floating-point scale degrees interpolated between semitones
Renardo's `.eclipse(dur, every, offset)` periodically silences a player for automatic arrangement breaks
Renardo's `linvar()` interpolates values linearly across its duration, unlike the step-wise `var()`
Renardo's `Pvar` makes entire Patterns switch over time, enabling structural variation between cycles
Renardo's `Ring` cycles through its elements indefinitely when called, unlike a one-shot list
Renardo's `var()` creates a time-varying value that switches between states at bar-aligned boundaries
Renardo's Ableton backend can automate parameters with TimeVar at up to 300Hz via AbletonOSC
Renardo's PRand, PChoice and PWhite generate stochastic values that re-roll each cycle
Renardo's REAPER backend controls instruments via OSC, enabling DAW-quality effects without SuperCollider
Renardo's TempoClock pre-queues events 0.25 seconds early to compensate for processing latency
Renardo's VRender extension converts note sequences and lyrics into a vocal WAV via sinsy.jp singing synthesis
Sardine embeds three distinct pattern mini-languages (Sardine PL, Ziffers, Vortex) selectable per player
SC Groups organize synth nodes into named sections that can be targeted for routing and order-of-execution control
SC receives MIDI via MIDIIn.connectAll and MIDIdef callbacks; ADSR gate management requires per-note synth node tracking
SC receives MIDI via MIDIIn.connectAll and MIDIdef, and tracks note on/off pairs using arrays indexed by MIDI note number
ServerBoot and ServerTree callbacks ensure SynthDefs and resources load in the correct order after server boot
Silence and restraint in a well-grooved set are active arrangement choices, not absence of ideas
Slight pitch offset between two identical synth voices creates beating and perceptual thickness
Slow coding is a live coding variant where extended periods of silence or stasis are made compositionally meaningful through visible anticipation
Some quality checks are permanently not-possible on this rig because the 4-bin FFT cannot supply the required feature
Sonic Pi _slide: opts cause synth parameters to glide smoothly between values over time
Sonic Pi auto-inserts incoming MIDI events into Time State, enabling sync and get on them
Sonic Pi filters sample packs by substring, regex, or symbol before selecting by index
Sonic Pi listens on port 4560 for OSC messages and routes them into Time State automatically
Sonic Pi sends MIDI messages in sync with music via midi, midi_note_on, and related functions
Sonic Pi's :tb303 synth emulates the Roland TB-303 with a dedicated cutoff envelope and resonance
Sonic Pi's randomness is deterministic so a 'random' pattern you like can be reproduced and performed
Sonic Pi's Time State (get/set) provides deterministic shared state across live_loops
Sonic Pi's Time State paths support OSC-style wildcard pattern matching for sync and get
Sonification maps non-audio data to sound parameters to convey information through the ear in SuperCollider
stack layers a list of patterns to play simultaneously so one transformation can shape them all
Staying 'in time' with the audience's phrase structure is the central discipline of live-coded dance music
Stochastic music controls broad statistical properties of a piece rather than specifying individual events exactly
striate cuts each sample into n equal grains, reorganising them for granular-texture effects
struct imposes a new rhythmic grid on a pattern while mask only silences it where the mask is 0
Subtractive synthesis removes frequencies from a rich source using resonant low-pass or high-pass filters
SuperCollider allows adding new methods to existing classes via extension files placed in the Extensions folder
SuperCollider classes and events both support object modeling: classes via inheritance, events via prototype chains
SuperCollider methods do not auto-return their last expression; the caret (^) designates the return value
SuperCollider Patterns schedule streams of synthesis events algorithmically using Pbind and combinators
SuperCollider's FFT UGens enable real-time spectral processing in the frequency domain
SuperCollider's functionality is extended via Quarks (language packages) and UGen plugins (server extensions)
SuperCollider's Pbind pattern system is comprehensive but hard to read and manipulate live, motivating constrained live-coding languages
SuperCollider's Pitch and Amplitude UGens extract frequency and loudness from audio input in real time
SuperCollider's RandSeed and RandID make stochastic synthesis reproducible from a given seed
sync in Sonic Pi always blocks until a future Time State event is inserted
Temporal recursion — a function calling itself with a time delay — is a core live coding time-representation technique (as in Extempore/Impromptu)
Tension is built by ramping one or two dimensions over a phrase then resolving on a boundary
The :slicer FX modulates amplitude rhythmically using a control wave at a configurable phase duration
The :slicer FX's probability: opt gates each phase on or off deterministically for random-sounding rhythms
The ~> and <~ operators shift a Tidal pattern forward or backward in time by a fraction of a cycle
The apparently simple 'reversal' operation has multiple non-equivalent implementations depending on assumptions about rests, events, and scale
The Cognitive Dimensions of Notation framework describes trade-offs in programming language design using named dimensions
The default livecoding agent mode is copilot: the human leads, the agent proposes and waits rather than acting autonomously
The halting problem means no algorithm can decide whether another will terminate, making perfect global repetition a sign of failure
The International Conference on Live Coding (ICLC) has been live coding's primary academic and community venue since 2015
The ixi lang matrix hybridises declarative score notation with full SuperCollider code per cell, enabling grid-based algorithmic navigation
The overarching taste check compares each intended intent-vector field against a perceived estimate within tolerance
The TOPLAP 'show us your screens' principle requires every livecoding change to land as a legible diff in the visible file
Tidal loops the current command until a new one compiles, with a completion rule ensuring musical continuity
Tidal pattern combinators transform time — not just events — enabling rotation, reversal, and every-nth-cycle transformations
Tidal provides continuous waveform functions (sine, tri, saw, square) that return smoothly varying values for use as effect parameters
Tidal represents time as rational numbers so musical subdivisions are stored exactly as fractions without floating-point error
Tidal schedules OSC messages to an external synthesiser, separating pattern logic from sound synthesis
Tidal's |>, <|, |<, >|, |+| operators combine two patterns by separately controlling which side provides structure vs. values
Tidal's begin and end play only part of a sample, and unit "c" makes speed stretch playback to fit the cycle
Tidal's cut stops any other sound using the same cut group, while legato controls how long a sound plays before the next one starts
Tidal's higher-order pattern transformations compose freely because they operate on functions, not on stored event lists
Tidal's pipe arithmetic operators (|+, |-, |*, |/) modify control values relative to an existing pattern value
Tidal's shuffle plays a random permutation of a pattern's parts (without replacement); scramble picks parts with replacement
Tidal's slice and splice chop a looping sample into n equal slices for rearrangement; splice also pitch-adjusts each slice to fit its step duration
Tidal's sometimes applies a function to random individual events; somecycles applies it to entire random cycles
Tidal's stripe repeats a pattern n times per cycle but with random sub-cycle durations that still fill the cycle
Tidal's wchoose picks from a list with weighted probabilities, unlike choose which weights all options equally
Tidal's weave function offsets multiple patterns in time while applying a shared effect pattern, creating canonic and spatial structures as side effects
TidalCycles `@N` sets an event's duration to N steps, enabling non-isochronous rhythms
TidalCycles `off` overlays a time-offset, transformed copy of a pattern to build canons
TidalCycles ascii, binary, and binaryN turn strings or integers into boolean rhythm patterns via their binary representation
TidalCycles bite slices a pattern cycle into equal-sized pieces and addresses them by index, enabling arbitrary reordering of pattern sections
TidalCycles can trigger custom SuperCollider synths using n with note names or midinote with MIDI numbers
TidalCycles choose emits continuous random values from a list; wchoose adds probability weights to make some values more likely
TidalCycles chunk divides a pattern into n sections and applies a transform to each in turn, one per cycle
TidalCycles degrade and degradeBy randomly remove events from a pattern at a controllable probability
TidalCycles euclidInv and euclidFull produce the rhythmic complement or dual-voice Euclidean pattern
TidalCycles loopFirst repeats only the first cycle of a pattern, freezing later cycles' content
TidalCycles mask gates a pattern with a boolean mask, letting events through only where the mask is true
TidalCycles necklace generates boolean rhythmic patterns from a list of inter-onset intervals rather than onset/total-steps parameters
TidalCycles palindrome alternates a pattern between forward and backward playback every other cycle
TidalCycles ply repeats each event in a pattern n times within its original time slot
TidalCycles range rescales a 0–1 continuous pattern to any numeric interval
TidalCycles ribbon cuts a fixed window from a pattern's timeline and loops that window
TidalCycles rot rotates the values of a pattern leftward while preserving the original rhythmic structure
TidalCycles spread applies a function with each of a list of parameter values in turn, cycling one value per cycle
TidalCycles stripe repeats a pattern n times at random speeds while keeping total duration constant
TidalCycles struct imposes a boolean rhythmic structure on any pattern, enabling Euclidean and binary rhythm shapes
TidalCycles stutter repeats each event n times separated by a fixed time offset, creating manual echo-style delays
TidalCycles timeLoop t loops a pattern's cycle sequence every t cycles, like a modulo on cycle number
TidalCycles trunc cuts a pattern short while linger repeats the kept fraction to fill the cycle
TidalCycles uses Ableton Link by default for BPM synchronization with other performers and applications
TidalCycles when and whenmod apply transforms based on a predicate or modular arithmetic test on the cycle number
TidalCycles' 'iter' rotates a pattern by a fraction each cycle, and the rotation amount is itself patternable
Two live_loops with slightly different sleep times produce Steve Reich-style phasing
Two pattern-transforming functions can be composed into one and applied together in TidalCycles
use_real_time removes Sonic Pi's scheduling look-ahead for MIDI and OSC-driven live loops
Vary the type of energy (density, brightness, space) not only the loudness level
When a craft choice conflicts with the livecoding ethos, the ethos wins — it is an operating constraint, not a footnote
When in doubt, an autonomous livecoding agent should wait rather than act
withValue (alias fmap) applies a function to every Hap's value without changing timing
G · Shaders & GPU programming — 61
A cheap bounding primitive tested first can skip the full SDF evaluation for the majority of ray steps
A compute shader is an @compute WGSL function that runs without fixed vertex/fragment I/O, reading and writing only storage buffers or textures
A for loop that repeatedly scales, tiles (fract), and accumulates color creates multi-scale fractal detail in shaders
A quadratic Bezier SDF takes three control points and returns distance to the curve
Adding abs() to a displaced SDF allows the raymarcher to backtrack when it overshoots a surface
An SDF can be derived from any implicit curve equation by setting the LHS equal to d
Animated SDF characters recover rest-pose coordinates via an invertible animation so textures track the surface rather than swimming
Animating the smooth-minimum blend radius with the motion signal adapts blending to the character's pose
Applying a smooth-step S-curve to output color increases contrast and creates a filmic look
Applying pow() with a value slightly above 1.0 to the final color enhances contrast by darkening shadows without affecting highlights
Audio FFT data can be passed to shaders as a 4-band uniform for audio-reactive visuals
BRDFs model how a surface reflects light as a function of incoming and outgoing directions
Colorizing the shadow penumbra with warm tones rather than grey makes rendered images more cinematic
Combining fast grid traversal with local SDF raymarching accelerates scenes where objects occupy sparse grid cells
Compute work in WebGPU runs in a compute pass that is recorded before the render pass in the same command encoder
dispatchWorkgroups takes the number of workgroups, not invocations — total invocations equals workgroup count times workgroup size
Dividing a shader build into named checkpoint stages lets you resume from a stable state and avoid rabbit holes
Exact SDFs give the true Euclidean distance; approximate SDFs are safe lower bounds but force smaller marching steps
Extending map() to return a vec4 provides auxiliary channels for UV coordinates, occlusion, and per-material signals
Fake subsurface scattering brightens silhouettes using the Fresnel term tinted by a flesh color
Fractional Brownian Motion adds correlated memory to white noise integration, controlled by the Hurst exponent H
GPU compute/ALU capacity growing faster than memory bandwidth made primitive-based SDFs increasingly competitive from 2007 onward
GPU ray tracing adds five programmable shader types to the graphics pipeline
Grouping scene elements rather than distributing them uniformly prevents visual noise and creates organic set design
Injecting fake bounce/subsurface lighting by hand into a raymarched scene fakes realistic results without global illumination
Making the raymarching hit threshold proportional to distance implements view-dependent level-of-detail
Microfacet models represent rough surfaces as collections of tiny facets described by a normal distribution
Modern real-time rendering uses rasterization + ray tracing + denoising together, not pure ray tracing
Monte Carlo integration converges at O(n^-1/2) regardless of problem dimension, unlike quadrature methods
Monte Carlo integration estimates the rendering-equation integral by averaging random samples
Motion blur is achieved by rendering multiple frames at sub-frame time offsets and averaging the results
Named noise colors (pink, brown, yellow) map to specific fBm spectral slopes and Hurst exponents
Noise-based fBm fills the frequency spectrum with far fewer octaves than sine-wave additive synthesis
Offline rendering's trajectory from rasterization to pure ray tracing predicts interactive graphics' future
Outdoor SDF shaders use three light sources: sun key light, sky dome fill, and ground bounce
Perlin noise is a repeatable pseudo-random function of 3D position used for procedural textures
Polynomial smooth-min functions preserve SDF shape outside the blend zone; exponential versions distort the whole field
Procedurally painted occlusion using model-aware distance signals fills gaps left by sampled AO
Radiance is constant along a ray through empty space, making it the natural quantity for ray tracing
Radiance is the fundamental radiometric quantity for rendering because it is constant along rays in empty space
Raising an animation curve to a power of itself creates a non-linear 'stay down longer' contact simulation
Raising per-surface occlusion to a power of the surface color simulates colorized secondary light bounces
Ray differentials estimate how much of a pixel's footprint overlaps SDF geometry to produce smooth antialiased edges
Raymarching surface normals are estimated by sampling the SDF gradient at nearby points
Real-time ray tracing is limited to ~1 ray per pixel; denoising reconstructs a clean image from noisy 1-sample results
SDF ambient occlusion samples the field at short offsets along the normal to detect nearby geometry
SDF soft shadows track how close the shadow ray comes to occluders, not just whether it hits them
Sky specular is computed by reflecting the view ray and checking if the reflection hits the sky
Spherical UV coordinates map latitude/longitude angles to texture image coordinates for sphere texturing
Starting with over-saturated colors and pulling back is more reliable than building up from grey
Storage buffers are large, compute-shader-writable GPU buffers declared var<storage> in WGSL, contrasting with size-limited read-only uniforms
The 1/x function creates neon glow effects in shaders by producing extreme brightness near zero and a slow falloff
The four core radiometric quantities for rendering are energy, flux, irradiance, and radiance
The fract() function repeats space by tiling UV coordinates into a 0–1 grid, enabling fractal-like layering
The Hurst exponent directly encodes the fBm's statistical self-similarity: zooming in by U horizontally scales amplitude by U^(-H)
The Phong model sums ambient, diffuse, and specular components to shade 3D surfaces
The ping-pong pattern uses two alternating state buffers — one read-input, one write-output — to prevent in-place mutation corruption in GPU simulations
WebGPU compute threads are grouped into workgroups and addressed by a global invocation id
WGSL storage buffers are read-only by default with var<storage>; read-write access requires var<storage, read_write> and is only available in compute shaders
When multiple pipelines share resources, an explicit GPUBindGroupLayout and GPUPipelineLayout must replace the auto-generated layout
Wrapping edge-cell neighbor lookups with the modulo operator creates a toroidal grid topology that prevents out-of-bounds buffer access
H · Creative coding & live-coded visuals — 63
2D Perlin noise maps x,y coordinates to smooth organic texture fields
A camera-screen feedback loop is a physical system that iterates a differential equation
A driver computes a property's value from other properties via a function or Python expression
A Hydra sketch with no .out() is guaranteed blank, so .out() presence is the [NOW] proxy for a black-frame check
A Hydra sketch with no time or reactive term is static by construction, the [NOW] proxy for a frozen-frame check
A p5.js filter shader reads the canvas via tex0 and vTexCoord to apply per-pixel post-processing
A small random deviation prevents emergent systems from collapsing into a uniform state
Braitenberg Vehicles show lifelike behaviour emerging from direct sensor-to-motor coupling
cables.gl operators are programmable in JavaScript with typed ports, extending the patch beyond the built-in op library
Cellular automata with continuous state values (0–255) produce fluid wave-like patterns
Chance operations use randomness as a deliberate artistic tool, not as an absence of control
CineCer0 live-codes video, image, and text as composable function chains in the browser
CineCer0's natural/every/round/chop functions align video playback to Estuary's cycle grid
CineCer0's ramp function linearly interpolates a style parameter over a given number of cycles
Circle packing fills a region by growing circles until they touch, then freezing each on collision
Circle-circle collision is detected by comparing center distance to sum of radii
Connecting nodes with springs that pull toward a rest length produces force-directed layouts and elastic motion
Conway's Game of Life produces biological patterns from two simple rules about neighbor count
Cross-dissolving two layers per pixel with lerpColor() and a coordinate-dependent wave makes non-uniform wave transitions
Defining a small set of behavioral rules for elements produces emergent visual complexity through their interactions
Diffusion-limited aggregation grows organic dendritic clusters by snapping each new particle onto the nearest existing one
Emergence is complex organized behavior arising from many simple local interactions
Envelope following maps smoothed audio amplitude over time to continuously control visual parameters
Feeding the previous frame back with an FFT-driven amount makes visual trails grow with the audio build
Fractals are shapes that exhibit self-similar structure across multiple scales
Frame differencing detects motion by comparing pixel values between consecutive video frames
Geometry Nodes builds geometry procedurally as a node tree flowing from Group Input to Group Output
Hydra can use other browser windows as live video sources via WebRTC peer-to-peer streaming
Hydra modulate functions use one source's colors to warp another source's geometry
Hydra video feedback is created by routing output into a named buffer and reading that buffer back
Hydra's loadScript() imports arbitrary JavaScript libraries (Three.js, Tone.js, p5.js) into the live editor
Image convolution applies a kernel matrix to each pixel's neighbourhood to produce blur and sharpen filters
Kinetic typography treats letterforms as visual elements that move, scale, and respond to input
Langton's Ant produces ordered structure from two purely local per-step rules
Lissajous figures are closed curves traced by combining two perpendicular sinusoids at integer frequency ratios
Mapping source brightness to tile size rasterizes an image into a halftone-like grid where dark areas make smaller tiles
Nodes that mutually repel each other self-organise into an even spatial distribution without explicit placement logic
p5.Framebuffer is an off-screen GPU surface you can draw to and then reuse as a texture
Parameterized randomness lets artists control how much chance enters a system via explicit ranges
Perlin noise mapped to angles at each grid cell creates a smooth flow field that steers particles or arrows organically
Points on a sphere's surface are computed from two angles using nested sin and cos
Processing class inheritance lets subclasses extend and override their parent's behaviour
Processing communicates with hardware boards over USB serial to read sensors and send control signals
Processing supports 3D via P3D and OPENGL renderers, specified as a third size() argument
Processing's blend() and filter() apply compositing modes and pixel filters to images
Processing's pixels[] array gives direct read-write access to every pixel of the display window
Providing a list of values to a Punctual shape distributes results across red, green, and blue channels
Ramping brightness down and blending toward a solid empties the canvas gracefully instead of a hard hush cut
Realistic flocking simulation requires only three local rules: separation, alignment, cohesion
Recursive draw functions produce fractal branching structures where each call draws one branch and spawns smaller sub-branches
Recursive fractal structures are coded as objects that instantiate child copies of themselves
Separating treemap layout calculation from rendering makes the algorithm reusable across any graphics library
Shadertoy fragment shaders run in three.js on a fullscreen quad by mapping iTime and iResolution uniforms into the render loop
Stepping kaleidoscope symmetry up on a downbeat gives a visual drop paired with an audio peak
Swapping the FFT index driving a visual parameter aligns the visual energy with the new section's audio driver
The ofxAssimpModelLoader addon loads and draws 3D model files in openFrameworks
The squarified treemap algorithm keeps each cell near-square by filling rows until adding the next item would worsen the aspect ratio
Three.js AnimationMixer plays and blends AnimationClips on a 3D object by being updated each frame with a delta time
Three.js post-processing chains render passes through an EffectComposer that processes them in order of addition
Toggling a color inversion or hard posterize on the drop downbeat is a single-frame visual hit
Treemap boxes can be styled with callback functions that receive value and index, enabling data-driven colour
Tying scroll or scale speed to a rising FFT band accelerates the visual motion through a build
Varying the initial spatial distribution of agents changes the global form that emerges from identical local rules
I · VJing, projection mapping, LED/DMX & video — 64
A 'last-touched' link system maps any MIDI or OSC input to any UI parameter without pre-assignment
A Cue List steps through a Chaser's scenes one cue at a time for theatrical operation
A LedFx virtual is a device-independent effect surface that can be spanned or copied across multiple physical LED strips
A replicator-plus-delay system distributes one visual engine across many screens by shifting the same signal in time per screen
A standard TOX container convention (out1 TOP + thumb TOP) makes VJ mixer cells interchangeable across projects
A universal TouchDesigner project template with pre-wired input/output buses reduces per-project configuration and enables flexible device swapping
Abstract AV visuals create a utopian space by giving entities behaviors that defy physics — metamorphosis, merging, spawning
Abstract visuals resist narrative analysis but can be composed and critiqued using the perceptual principles of music
Abstraction in live AV defeats the audience's conditioned expectation to read images as narrative
An EFX function generates mathematical movement paths on pan/tilt or RGB channels for automated fixture motion
An RGB Matrix function projects patterns and text onto a grid of RGB fixture heads, scriptable in JavaScript
Audience participation in live cinema spans a spectrum from passive watching to active co-creation of the visual environment
AV improvisation between visualist and musician requires real dialogue; in practice the visual artist often follows rather than co-creates
Centralising all sensor inputs through a single OSC/WebSocket bridge (MAMI pattern) lets TouchDesigner focus on interaction design rather than device plumbing
Channel Modifier curves remap DMX output values non-linearly, enabling LED linearisation and channel parking
Compressing inactive TOP networks to one-eighth resolution recovers GPU memory headroom in multi-scene VJ rigs
Digital physics simulations (rigid body, fluid, cloth) are effective for interactive installations because humans are calibrated to physics-based motion
Effective live visual sets balance continuity (flow) and surprise (break): too little novelty is boring, too much destroys coherence
Fixture Remapping reassigns an existing project's channels to different fixtures in minutes
Gestural interfaces require a visible, immediate correspondence between the performer's movements and the audiovisual output
In live cinema, software is used as a real-time instrument for expression, not as a tool to produce a finished product
Inverting a mask on a black solid top layer turns it into a cutout window onto lower layers
Linear Light combines Linear Dodge and Linear Burn and simplifies to base + 2·top − 1, decreasing contrast
Live cinema is a hybrid of opera's theatrical gesture, painting's intimate craft, and the concert hall's performative presence
Live cinema is disputed between an abstract-poetic model (Makela) and a narrative-storytelling model (Harris)
Mapping a per-track OSC envelope to visual parameters turns a music sequencer into a precise visual accent machine
NDI and Spout outputs let a TouchDesigner rig feed any downstream software without a physical video signal
Node-and-wire diagrams are not just tools but a diagrammatic aesthetic that restructures how an artist perceives and generates form
Paracinema identifies cinematic properties realized in non-filmic media, without the standard film apparatus
Parameterising phase offset across objects produces overlapping-action animation generatively without per-object keyframing
QLC+ exposes a web interface for headless remote control of the Virtual Console and Simple Desk over HTTP
QLC+ kiosk mode locks the UI to Virtual Console only, preventing editing during live operation
QLC+ maps MIDI beat clock start/stop and beat signals to special channels 530–531 for BPM-sync
QLC+ RGB Scripts are self-executing JavaScript objects implementing rgbMapStepCount and rgbMap functions
Restarting several clips together in VPT needs per-source trigger buttons or a shared router controller
Source presets store media state separately from VPT presets, so mapping and content recall independently
Stage visibility allows audiences to distinguish live VJ performance from pre-recorded playback and validates the VJ as a performer
Structuring a TouchDesigner performance rig as Control, World, Core, and UI containers separates concerns and speeds iteration
Synaesthesia motivates early visual music but is no longer its defining criterion
TDAbleton enables bidirectional MIDI and audio-feature communication between TouchDesigner and Ableton Live, allowing visuals to drive audio and audio to drive visuals
The 'visual wallpaper' problem arises when VJs compete for attention rather than complementing the music and space
The Audio Trigger widget maps audio spectrum bands and volume to DMX channels, functions, or VC widgets
The club context demands atmosphere over narrative; the gallery or theater sustains focused attention — context determines appropriate AV strategies
The Loopback plugin routes QLC+ DMX output back as input, enabling Scenes to control Virtual Console sliders
The mesh editor extends corner-pin mapping to a grid of draggable points for curved surfaces
The QLC+ Script function automates sequences of function start/stop and DMX channel commands with wait times
The Show Manager places QLC+ functions on a multitrack timeline for time-driven light shows
The surprise a performer feels discovering a visual effect in real time is not automatically shared by the audience
The XY Pad widget controls moving head pan/tilt with mouse, keyboard, or external controller and supports presets
Turning off cooking for invisible TOPs eliminates wasted GPU time in multi-layer VJ rigs
Universe Passthrough forwards external DMX input directly to output, enabling protocol conversion and data merging
Vivid Light combines Color Dodge and Color Burn around middle grey, increasing perceived contrast
VJ skills transfer beyond nightclubs into installation art and immersive architectural performance
VPT emits OSC loopreport and cliptime events so external tools can sync to video progress
VPT exchanges sensor and actuator data with Arduino over serial using an id-plus-value format
VPT exposes its full parameter set as typed OSC commands across sources, layers, presets and control
VPT's built-in LFOs automate any routed parameter with sine, ramp, triangle or square waves
VPT's cuelist is a text script of one-letter commands that sequences presets and parameters
VPT's mix module composites two sources with a blend mode before the result feeds a layer
VPT's router maps a numbered controller to any parameter via rows keyed by controller number
VPT's SoundTrigger app maps left/right audio channels to VPT controllers for audio-reactive visuals
VPT's Video Trigger detects motion in camera zones by background subtraction to fire events
WLED supports multiple named ledmaps selectable per preset
Youngblood defines expanded cinema as expanded consciousness, not a specific set of film technologies
J · Audio-visual integration & sync — 53
A DAW's internal musical state, not just its audio, can drive visuals when exposed over OSC
A jitter buffer absorbs packet timing variation by introducing a small constant delay before playback
A jitter buffer must define policies for underflow (empty) and overflow (full) conditions
A MIDI 2.0 device must implement MIDI-CI-with-discovery plus a feature, or UMP-with-discovery plus a feature
A MIDI-CI Profile is a published set of messages and required responses that devices can advertise and negotiate
A performer's live audio plus tracked position can drive visuals anchored to their body in real time
Ableton Link lets visual software lock video clips and generative patches to a shared musical clock
AbletonOSC exposes per-track output meter levels, enabling audio-reactive visuals driven by individual track energy
AbletonOSC implements OSC wildcard dispatch by converting * to a [^/]+ regex and matching all registered handlers
AbletonOSC lets you add, remove, and query MIDI notes in a clip programmatically over OSC
AbletonOSC supports OSC wildcard patterns using * to query all sub-addresses matching a path segment
AbletonOSC's Device API lets you read and write any synthesiser or effect parameter by index via OSC
AbletonOSC's MidiMap API creates persistent MIDI CC assignments to device parameters via OSC, without Max for Live
AbletonOSC's OSC server uses a non-blocking UDP socket to drain all queued messages per tick without blocking Live
AbletonOSC's property listener uses Live's add_<prop>_listener API and immediately sends the current value on subscribe
Adaptive jitter buffers resize dynamically to stay as small as possible while preventing dropouts
An Ableton Link session does not require Ableton Live to be part of it
An audio ADSR envelope and a visual physics simulation share the same integrate-over-time structure — both accumulate an impulse into a changing state
An OSC bundle groups multiple messages with a shared timetag so they are dispatched atomically at the same time
Analysing sound features (frequency, tempo, gesture) as inputs to probabilistic rules generates visuals that respond to music without one-to-one mapping
Any numeric parameter in Hydra can be replaced with a function, enabling audio or MIDI to modulate visual parameters
Autocorrelation detects pitch by finding the lag at which a signal most resembles itself
Automation and Clip Envelopes in Ableton Live can generate discrete visual output cues without audio processing
AV sync ranges from fine-grained per-parameter control for soloists to coarse cue-passing for visualist collaborators
Bidirectional communication is the core innovation of MIDI 2.0, letting devices discover and auto-configure for each other
Each Live scene can carry its own tempo and time signature, settable via OSC
Gibber maps audio objects directly to visual properties using bias and scalar for range control
Jitter buffers use per-packet timestamps to reorder out-of-sequence audio packets
Jitter is variation in latency over time, not latency itself
Link keeps apps in time by relating independent timelines, not by forcing one shared timeline
Link start/stop synchronization only follows explicit user actions and is not auto-applied when joining a session
Live's embedded Python cannot use threads; AbletonOSC polls the OSC socket once per 100ms tick using Live's scheduler
Logarithmic and 1/3-octave FFT scaling match frequency display to human pitch perception
Many self-timed audio voices and many self-timed visual agents are the same concurrent-agents structure expressed in two domains
MIDI 2.0 extends MIDI 1.0 rather than replacing it, so existing devices and messages stay valid
MIDI-CI is MIDI 2.0's discovery-and-negotiation layer, delivering profiles, property exchange, and process inquiry
MIDI-CI Property Exchange uses JSON over SysEx to get and set device data with no bespoke software
MIDI-CI runs over any MIDI 1.0 transport because it is carried in System Exclusive messages
MIR uses content-based audio feature extraction and machine learning to automate musical analysis and similarity search
Onset detection locates the start of new sound events in an audio signal by finding rapid increases in energy or spectral flux
Pre-rendered audio analysis APIs supply timestamped beats and pitch for offline sync
Robust beat detection uses a decaying cutoff plus a hold window to debounce triggers
Splitting incoming audio into named bands lets a VJ hook each band to a different visual parameter
Stretching a video clip to bar length couples visual rhythm directly to musical tempo
The /live/song/get/track_data command queries multiple tracks and clips in a single OSC message
The AbletonOSC Python client uses a threading.Event to block on an OSC reply, creating a synchronous query interface
The Link quantum value sets the loop length for phase synchronization, aligning loop boundaries across apps
The MIDI 2.0 Protocol raises resolution on all Channel Voice Messages and adds per-note expression
The strongest section changes fire an audio and a visual transition on the same downbeat
The track_data handler uses a dot-namespace format to query track, clip, clip_slot, or device properties in a single call
The Universal MIDI Packet is one container for all MIDI, adding 16 Groups of 16 channels (256 total) and jitter-reduction timestamps
Two MIDI 2.0 devices reach full communication through an ordered discovery sequence ending in normal MIDI use
Vision and hearing process time and density differently — AV artists must account for these asymmetries to achieve coherence
K · AI & real-time generative AV — 101
A ComfyUI custom node is a Python class with INPUT_TYPES, RETURN_TYPES, FUNCTION, and CATEGORY
A DDSP ProcessorGroup chains processors as a DAG, configurable via gin without Python code changes
A deep source separation model can be wrapped in nn~ to split live audio into stems in real time
A GrooveTransformer ML model can predict a full drum pattern from a sparse hit/velocity/offset input vector
A screen region or webcam feed can be the continuous image source for real-time img2img diffusion
A warmup phase fills the pipeline's latent buffer and JIT-compiles kernels before timing starts
An autoencoder compresses input to a latent code whose interpolations generate smooth variations
Angular cumsum avoids phase drift in long synthesis by chunking cumulative sums and resetting with modular arithmetic
Any PyTorch model can become a nn~ object by subclassing nn_tilde.Module and registering methods and attributes
Arithmetic on a GAN's latent vectors edits generated images semantically
Averaging predictions over multiple time-shifted inputs (the shift trick) improves separation quality
Classifier-free guidance combines conditional and unconditional model outputs at inference to steer generation without a separate classifier
Classifier-guided diffusion steers generation by adding classifier-score gradients to the noise prediction
ComfyUI's HTTP API accepts workflows as JSON with nodes keyed by numeric IDs and links as [node_id, slot_index] arrays
Compiling UNet and VAE to TensorRT engines delivers the largest throughput gain in StreamDiffusion
ControlNet adds spatial conditioning to a frozen diffusion U-Net via a trainable copy connected with zero-convolution layers
DDIM makes diffusion sampling deterministic by setting the stochasticity parameter eta to zero, enabling far fewer steps
DDPM keeps the reverse-process variance fixed and only learns the mean, simplifying the training target
DDPM ResNet blocks inject the timestep embedding via scale-and-shift (FiLM-style) conditioning
DDPM's training objective is to predict the noise added to an image rather than the original image directly
DDSP timbre transfer re-synthesizes audio from one instrument using a model trained on a different instrument
DDSP's frequency_filter designs FIR filters from frequency-domain magnitude curves using IRFFT and windowing
DDSP's RnnFcDecoder runs independent FC stacks per conditioning input, concatenates, then runs an RNN
DDSP's Wavetable synthesizer reads a learned single-cycle waveform at a time-varying phase
Deforum's 3D FOV scales how fast translation_z moves the canvas, with defined edge cases at 0, 180 and negative values
Deforum's anti-blur applies an unsharp mask to counteract the progressive blurring that builds during long animations
Deforum's Perlin noise injection adds organic, spatially coherent variation to frames rather than uniform random noise
Demucs exposes a Python API for integrating stem separation into scripts and pipelines
Demucs uses a weighted ensemble (BagOfModels) of individually trained checkpoints for best performance
Demucs' internal separation output is a 4-D tensor shaped [batch, sources, channels, time]
Diffusion models are attractive because they are simultaneously analytically tractable and flexible
Diffusion models learn to reverse a Markov chain that progressively corrupts data with Gaussian noise
Diffusion models trade sampling speed for stable training and broad mode coverage relative to GANs and VAEs
Exponential Moving Average of weights produces smoother RAVE models by averaging checkpoints over time
Extending stem separation to guitar and piano is harder than the standard four-stem split
Feature matching loss in RAVE aligns intermediate discriminator activations between real and generated audio
Frequencies softmax maps network logits to Hz by weighting a log-spaced frequency grid
Fusing LCM-LoRA into an SD model enables 2–4 step diffusion inference without retraining the base model
Gin's @gin.register limits side-effects by only injecting defaults when a function is used as an argument
Group normalization divides channels into groups and normalizes within each group, working well with small batch sizes unlike batch normalization
Hybrid source separation processes audio in both waveform and spectrogram domains simultaneously
img2img and txt2img mode in StreamDiffusion differ in their latent initialization and CFG constraints
in_ratio and out_ratio in nn~ method registration define the temporal compression between audio and model outputs
Key DDPM follow-up works include improved variance learning, cascaded generation, classifier guidance, and classifier-free guidance
Linear attention scales O(n) in sequence length — making it practical for image feature maps without the O(n^2) cost of full attention
Manipulating individual latent dimensions of a RAVE model morphs continuously between audio-effect and synthesizer behavior
mc.nn~ processes multiple audio channels through one RAVE model instance to cut CPU and RAM
mc.nn~ runs several audio streams through one model as a batch, saving CPU and RAM versus duplicating nn~
mcs.nn~ packs all of one instance's inlets/outlets into a single multi-channel connection, enabling per-batch latent operations
Multi-scale SpectralLoss compares audio at multiple FFT sizes to balance time and frequency resolution
Multiple LoRA weights can be loaded and fused into StreamDiffusion before inference for style mixing
MUSDB-HQ is the standard benchmark dataset for music source separation research
nn~ can output audio-analysis features as control-rate signals by using a large out_ratio
nn~ is a Max/PureData external that bridges trained neural audio models (RAVE, vschaos2) into a patching environment
nn~'s void/lazy mode lets you fix inlet/outlet count before attaching a model
Pre-computing KV-cache for the text prompt eliminates repeated text-encoder work across frames
Quantization maps high-precision weights to lower-precision formats using a scaling factor to preserve dynamic range
Raising RAVE's discriminator update period fixes phase-2 instability when the discriminator is too strong
RAVE architecture configs trade reconstruction quality, GPU memory, and control modality
RAVE can be exported to ONNX format for deployment in environments that do not support TorchScript
RAVE ramps the KL regularization weight over a warmup schedule to avoid posterior collapse
RAVE trains in two sequential phases: reconstruction first, adversarial refinement second
RAVE uses Pseudo Quadrature Mirror Filters to split audio into sub-bands before encoding
RAVE v3 uses Adaptive Instance Normalization to transfer timbre from one audio stream to another
RAVE's causal convolution mode lowers latency at the cost of quality by removing future context
RAVE's discrete mode quantizes latent vectors using Residual Vector Quantization
RAVE's exported latent size is chosen by a fidelity threshold on the PCA explained-variance curve
RAVE's phase 1 length is a fixed step count, not a quality-based stopping criterion
RAVE's temporal receptive field sets the minimum audio context and chunk length it can process
RAVE's variational encoder reparametrizes latent samples using the mean and log-variance trick
Replacing the standard VAE with TinyVAE (TAESD) reduces encode/decode latency at small quality cost
Residual CFG approximates classifier-free guidance at near-zero extra cost by recycling a stored noise residual
SDR (signal-to-distortion ratio) is the standard metric for evaluating stem separation quality
Separating image generation and display into different OS processes prevents Python's GIL from throttling GPU throughput
Splitting RAVE encode and decode in nn~ lets performers process individual latent dimensions live
Stable-Fast compiles the full diffusion pipeline with Triton + CUDA graphs for an intermediate speed tier
StreamDiffusion achieves 106 fps txt2img and 93 fps img2img on RTX 4090 with SD-Turbo plus TensorRT
StreamDiffusion achieves real-time AI image generation by batching denoising steps and skipping redundant frames
StreamDiffusion can process offline video frame-by-frame with img2img to produce a stylised output video
StreamDiffusion requires CUDA GPU, Python 3.10, PyTorch 2.1, and optional TensorRT for optimal performance
StreamDiffusion's IO queues decouple the input capture rate from the diffusion throughput via multiprocessing
StreamDiffusion's stream batch runs multiple denoising steps in parallel to sustain real-time frame rates
StreamDiffusionWrapper provides a high-level production interface; the raw StreamDiffusion class gives full control
t-SNE embeds high-dimensional feature vectors in 2D so perceptually similar items cluster together
TAESD provides fast high-quality latent previews without running the full VAE
Text conditioning is injected into the UNet by attention layers placed between ResNet blocks
The 'nice property' lets you sample any diffusion timestep directly from the original image without iterating the forward process
The DDPM ELBO decomposes into per-timestep KL divergences between Gaussians, each computable as an L2 loss
The DDPM U-Net assembles encoder, bottleneck, and decoder stages using ModuleList, with skip connections via concatenation
The DDSP autoencoder encodes loudness and f0 plus a learned latent z, then decodes to synthesizer parameters
The denoising U-Net uses sinusoidal position embeddings to communicate the current noise level (timestep) to every layer
The forward diffusion process adds scheduled Gaussian noise over T steps until the signal becomes isotropic noise
The forward method runs a neural model as an audio effect: audio in, neural-transformed audio out
The reverse diffusion process is a neural network that approximates the intractable denoising posterior at each timestep
The Stochastic Similarity Filter skips GPU work probabilistically when consecutive frames are nearly identical
The StreamDiffusionTD operator wraps StreamDiffusion as a TouchDesigner node for diffusion-based real-time visuals inside a TD network
The t_index_list selects which diffusion timesteps to apply, controlling quality–speed trade-off
TouchDesigner can host real-time AI/ML pipelines for generation, tracking, and stylisation in a live AV rig
U-Net's symmetric downsampling-upsampling structure with skip connections makes it a standard diffusion backbone
unCLIP generates images by mapping text to a CLIP image embedding, then decoding that embedding to pixels with diffusion
xformers memory-efficient attention reduces VRAM usage in diffusion UNets at modest speed cost
L · Visual foundations — 14
Appeal is the charm and charisma that makes an animated subject engaging to watch
Each hue has a perceived brightness (luminosity) that peaks at yellow/cyan/magenta and dips at red/green/blue
Exaggeration pushes an action beyond the literal while staying true to reality
Kandinsky proposes three primary contrasting pairs linking line type, plane shape, and color: straight/triangle/yellow, curved/circle/blue
Lines have color analogies: horizontal ↔ black, vertical ↔ white, diagonal ↔ red, free lines ↔ yellow and blue
Objective color principles override personal taste when the context has fixed requirements — meat markets, confectioneries, and floral occasions all specify palette constraints
Recreating a master's color palette in cut paper reveals its color instrumentation — relationships rather than specific pigments
Secondary action is an additional motion that reinforces and adds dimension to the main action
Simultaneous contrast can be neutralized by adding the missing complementary explicitly, or by introducing light-dark contrast between the hues
Solid drawing treats forms as three-dimensional objects with volume and weight
The color occupying the larger area in a composition acts as background; the smaller area advances — and these roles can reverse as area proportions shift
The expressive meaning of each color is the complement of its complementary's meaning — mixed colors inherit blended meanings
The eye spontaneously groups scattered same-color areas into a visible shape — these 'simultaneous patterns' are independent organizational elements in composition
The two diagonals of the basic plane carry contrasting tensions: the lyric diagonal (calm) and the dramatic diagonal (complex)
M · Performance, DJing, live-set & stagecraft — 48
'Ringing out' a monitor system identifies feedback frequencies with EQ to maximize gain before feedback
A 1/3-octave graphic EQ has Q≈4.31, not the 100 that engineers often assume
A commercially released mix tracklist is only final once every track clears licensing
A frenchcore DJ set ranges from 180 to 220+ BPM, sometimes closing with terror or speedcore
A heavy beat layered underneath a heavily tempo-shifted track masks its distortion artefacts
A microphone's polar pattern is non-uniform front-to-back, making rear-pointing feedback tests inaccurate
A performer who only plays their classics ends up permanently measured against their own peak
A semitone rise between tracks creates an energy boost by adding 7 (one semitone) or 2 (two semitones) to the Camelot number
A short loop on a shared chord bridges two harmonically compatible but chord-order-incompatible tracks
A sound engineer's core job on a DnB rig is keeping DJ music and MC mic at equal, non-masking levels
A stage monitor system is a separate sound system pointed at performers, providing individualized mixes
A sustained all-night dance floor is best held around 133 BPM, with faster bursts risking losing the crowd
All-hardware live sets require regular rehearsal and accept reliability constraints in exchange for genuine real-time improvisation
An absolute major/minor key switch keeps the root note and shifts Camelot number by ±3
At very high BPMs, long seamless blends work against the music; short cuts and splices serve the energy better
Camelot wheel major-to-minor harmonic mixing matches the number and changes the letter (A/B)
Chopping and blending are two DJ mixing strategies with opposite tradeoffs for following the music
Completing a production requires an immersive lock-in, not piecemeal sessions
Cutting a dubplate imposes an economic discipline that forces honest quality evaluation before release
Echoic memory lasts only ~20 seconds, making accurate cross-room comparison unreliable
EQ swaps do not fix key clashes — mismatched keys still clash after a bassline swap
EQ treats speaker-to-room interactions — it cannot fix fundamental room acoustic problems
IEM stage monitoring should average no more than 95–97 dB for multi-hour performances
In sound clash competition, record selection decides the win over technical mixing prowess
Layering means recording the rhythmic bed first, then overdubbing melodic and textural parts on top
Mastering engineers apply objective technical skills but aesthetic outcomes vary — choose one whose taste aligns with yours
Musicians can recalibrate their hearing to lower IEM levels within a few weeks of consistent practice
PA tuning begins with listening to known references, not measuring — verify before correcting
Pairing genres with shared cultural lineages creates combinations that feel coherent even when they sound unexpected
Placing an exceptional track between average ones amplifies its impact through the contrast effect
Playing at slow tempos exposes the quality of individual sounds because sparse arrangement gives each element nowhere to hide
Playing dubplates for a year before vinyl release uses live audience response as A&R
Playing records at the wrong RPM (33 instead of 45, or heavily pitch-shifted) reveals hidden qualities and creates new textures
Reacting to the crowd in real time — rather than following a pre-selected playlist — produces better dancefloor outcomes
Reading crowd composition — gender ratio, age, energy — shapes the entire arc of a DJ set
Reading crowd signals and adjusting song selection in real time is a core live DJ skill
Recording and reviewing your sets calibrates whether mistakes were as bad as they felt
Removing one IEM during a performance doubles hearing risk by exposing one ear while cranking the other
Repetition without change relaxes the listener into the rhythm, triggering excitement through predictability rather than surprise
Saving a reference render at the end of each session and reviewing it fresh before the next prevents perspective loss during deep detail work
Staying present-focused rather than planning career trajectories keeps a performer responsive to what is actually happening
Stopping the music briefly raises crowd anticipation more reliably than continuous sound because silence signals an imminent change
Sustained rhythmic consistency induces a hypnotic dancefloor state regardless of the absolute tempo
Tempo does not equal intensity: syncopation and arrangement can make slow tracks feel fast and fast tracks feel slow
Tone matching EQ corrects spectral mismatches between tracks from different genres or eras
Turning the volume down at the right moment is one of the most powerful crowd-manipulating effects available to a DJ
Use parametric EQ on mic groups for feedback control — bypass filters until needed at soundcheck
Warning a crowd in advance is a tool for surviving the expectation gap when you change genre
N · Tools & free/open software stack — 26
A clipped amplifier outputs approximately double its rated continuous power, threatening loudspeaker voice coils
A ground loop forms when two pieces of equipment share multiple ground paths, creating a hum-inducing loop antenna
Ableton's warp modes use different algorithms matched to the audio material being time-stretched
An active crossover placed before power amplifiers increases headroom, damping, and reduces distortion versus passive crossovers
An audio transformer provides galvanic isolation and common-mode rejection to break ground loops
Ardour's Cue window enables clip-launching and non-linear performance workflow alongside the timeline
Ardour's Strict I/O mode enforces matching channel counts through the plugin chain; Flexible I/O allows width changes
Audio plugin parameters can be defined in a CSV schema and the parameter boilerplate auto-generated to reduce repetition
Bridging a stereo amplifier combines both channels in mono to roughly double the power output
Clip envelopes automate parameters only within their clip while track automation applies across the full arrangement
Constant-voltage (70V) distribution allows many loudspeakers to share one amplifier using impedance-transforming taps
Constrained, primitive gear setups can lend a charm that more capable technology removes
Exposing a WebSockets and OSC interface lets any UI — web page, hardware controller, script — control an audio plugin at runtime
Focusrite Scarlett XLR inputs cannot take line-level signals; line sources must use the combo TRS input instead
JUCE's ValueTree structure can serve as a single source of truth for a plugin's complete state, driving all object creation
Loudspeaker polarity reversal is not the same as phase — polarity is frequency-independent
Loudspeaker sensitivity (dB SPL / 1W / 1m) plus power ratio in dB gives maximum SPL at 1m
Music-centred design extends user-centred design to privilege the needs of performers, audiences, and musical flow — not just usability metrics
Note FX process MIDI note data before the instrument, so a fixed clip can vary at performance time without editing it
Operate a live sound system 6 dB below the feedback threshold to maintain stable gain
Pure Data is extended with community externals installed via the Deken package manager
SOURCE presets are XML files where PRESET > SOUND > SOUND_SAMPLE hierarchy mirrors the sampler's class structure
SOURCE separates audio engine, sensor I/O, and API communication into three independent processes connected by OSC/WebSockets
Speech reinforcement needs 70–80 dBA; music reinforcement requires 85–100 dBA with 10–20 dB headroom
The Grid is Bitwig's patchable modular environment for building custom synths and effects inside the DAW
Troubleshoot 'no signal' by tracing from source to output; troubleshoot 'unwanted signal' by tracing from output to source
O · Culture, history & theory — 84
A 2001–02 Detroit trip exposed footwork producers to a faster, more polished ghetto-tech scene and pushed them toward a radio-friendly sound
A descriptive score documents what was heard; a prescriptive score directs a performer
Adrian Sherwood's On-U Sound applied dub production logic to industrial music, forging a rare Afrofuturist-industrial crossover
Afrofuturism reclaims racial otherness by recasting disenfranchisement as alien, using technology to become 'out of this world'
Ambient composition can be used as a vehicle for political and identity critique, not only as apolitical background texture
An underground scene's cultural impact often becomes legible only after it has dissolved
Bass-centricity can be a through-line across genre changes, enabling an artist to shift style while maintaining sonic identity
Big beat established templates for arena-scale electronic music that later genres (brostep, EDM) inherited
Big beat's decline was caused by overexposure through licensing, rising cocaine culture, and creative stagnation
Bristol's Purple sound (2008) fused dubstep with 1980s synth-funk and G-funk, seeding future bass
By the 2010s nu-disco production had permeated mainstream pop, with producers often working anonymously for pop acts
Chicago footwork battles use randomly chosen neutral judges from outside the competing crews to prevent favoritism
Colour bass combines brostep's impact with melodic dubstep's rich tonality in vibrant mid-range sound design
Crowd literacy for complex breakbeats is cyclical: lost during simple 'rolling' eras, regained when producers challenge it
Data sonification requires inference-preserving mappings: conclusions drawable from the sound must correspond to conclusions about the source data
Détournement recontextualises an existing sign to turn its own authority against itself
DJ Fulltono of Osaka pioneered Japanese footwork from 2008, persisting past crowds who found the tempo too fast
DnB distribution shifted from 12-inch vinyl singles to digital download to streaming, tracking the wider EDM market
DnB producers like Droppin' Science used polyrhythmic breakbeats to produce kinesthetic, bodily responses rather than purely sonic ones
DnB with 'flavour' (musical identity, replay value) outlasts technically-competent but disposable tracks
Drexciya extended electro's afrofuturism into a sustained science-fiction aquatic mythology
Drum and bass is dominated by independent labels run by DJ-producers, which maintained genre control outside major labels until 2016
Drumfunk foregrounds rhythmic complexity of breakbeats by minimizing bass and melody
Dub techno uses only a few elements filtered down low and sparse, filling the space with delays and reverbs
EBM and industrial draw on pre-modern British Isles ballad repetition patterns, not only African-diaspora groove traditions
EBM's 'body music' ideology demanded an aggressive physical performance style, rejecting the static synthesizer act
EBM's use of totalitarian and military imagery is an ironic-provocative strategy inherited from punk, not political endorsement
Effective chance in art is never blind — it is planned, constrained, and then surprising within those constraints
Electroclash's core stance was double-coded — critiquing the excess it simultaneously celebrated
Ethical field recording of indigenous communities requires giving tangible benefit back to the source
First-wave Detroit techno came from the suburban Black middle class, carrying a class tension into its aesthetics
Footwork producers treat any genre as valid input as long as the footwork rhythmic grid and bass drive the track
Footwork producers typically build a track by cutting up samples first and working the beat around them, at very high output rates
Footwork spread to Latin America in the late 2010s, with Mexico's JukeMX blending it with Latin percussion and baile funk
Footwork's signature emerged from removing the bass kicks and replacing claps with snares and hi-hats
French house splits into a space-disco strand, a Euro-disco-update strand, and a deep-American-house strand
Genre labels create implicit rules that constrain what producers feel they can make within a scene
Genres pass through four stages from avant-garde to scene-based to industry-based to traditionalist
Glitch music treats digital malfunction as critique of the myth of technological perfection
Glitch pieces strip electronic music to minimum information — atomic, anechoic, 1-3 minute durations — inverting electronica's layering ethos
Hip-hop and industrial converged on the sampler as a shared political tool at the same historical moment
House and techno's affective power draws on the Black gospel tradition of repetition and collective trance
Hypnotic techno achieves trance by stripping arrangement to expose repetition rather than adding elements
In DnB production, maximum rhythmic complexity can coexist with extreme surface minimalism — 'minimal-is-maximalist'
In dub techno, live-recorded automation of filter cutoff and send levels does the work of arrangement
In industrial music, noise functions as an emancipatory strategy to overload and disrupt perceptions of order
Industrial music structurally relies on extremist imagery rather than merely tolerating it
Industrial music sustains a permanent tension between pan-revolutionary ideology and danceable pop appeal
Industrial music uses cut-up and détournement to recycle existing media fragments without concealing the seams
Industrial techno's 2010s revival drew a post-dubstep audience but faced 'sounds old' criticism
Industrial vocals are traditionally distorted to reject the sonic clarity that marks authority
Introducing new elements every 8 bars and gradually increasing drum complexity sustains energy throughout a progressive house track
Jacques Attali argued that noise in music prefigures rather than reflects social transformation
Japan built one of footwork's most vibrant international scenes, adapting it into subgenres like vocaloid juke and party-juke
Jersey club seeded the deconstructed club movement in the early-mid 2010s NYC underground
Jersey club's mainstream adoption raised concerns about outsiders coopting the sound for bookings
Live coding deautomatizes algorithmic processes for audiences by making the generating code visible
Mainstream crossover can dilute an underground scene's credibility and hasten its decline
Mainstream pop absorbs niche LGBTQ+/women-led aesthetics while stripping their radical edge
Mainstream trance was criticized for following a fixed pop-format structure
Major label involvement in Grime diluted the genre by remaking artists into pop acts rather than amplifying the existing sound
Minimal techno creates groove by nesting rhythmic layers inside each other rather than layering sounds
Mirror neurons fire both when performing and when observing an action, enabling embodied perception of musical effort
Modern New Beat (2010s) is more futuristic and cinematic than Old New Beat and directly precedes Midtempo Bass
Neurofunk basslines twist and morph through complex modulation to a menacing, robotic timbre
Never tell someone your work is great before they hear it — let the work speak
New Beat is a mid-tempo (90–120 BPM) Belgian subgenre of EBM
Norwegian 'Norse House' fused space disco with nu-disco into a cosmic sub-scene led by Lindstrøm, Prins Thomas and Todd Terje
Planet Mu's Bangs and Works compilations (2010–11) and Hyperdub broke footwork to an international audience
Producing music requires being genuinely inspired rather than replicating a past sound — forced genre consistency produces hollow work
Progressive house tracks follow an intro–verse–build-up–drop–breakdown–outro structure in 8-bar segments
Progressive house transitions use white noise sweeps, filter rises, pitch risers, and reverse effects to bridge sections
RP Boo keeps the Roland R-70's analog warmth because digital transfer loses the punch that defines footwork's bass
Sambass is a Brazilian DnB fusion that incorporated samba and bossa nova rhythmic elements into the genre
Sidechain compression keyed to the kick creates progressive house's signature pumping bassline effect
Software emulation of hardware instruments adds representational layers that distance the user from the original circuitry
Strings of Life was performed live on a keyboard, not sequenced — it was a real-time production
Techno's mainstream commercial absorption lagged its underground creation by roughly 15 years
Technology lowers the barrier to making sound but raises the risk of producing without musical development
The demoscene has functioned as a training ground and talent pipeline for the European games industry
Underground Resistance emerged as a second-wave Detroit techno force defined by militancy, mystery, and sonic warfare
Visually uncompromising music videos can define how a genre's audience understands and remembers its music
West Side Chicago production favored sample-free rhythmic 'beat tracks', unlike the sample-based South Side style
What separates aspiring producers from releasing is often confidence, not technical skill
P · Community, scenes & practice — 9
A distinctive, collectible visual identity gives a label the recognizable mystique that a compelling sound alone cannot deliver
A live-coding community's code of conduct names domain-specific failure modes beyond generic anti-harassment rules
A per-release marketing plan fixes messaging, timing, and platform actions before a track goes live
Adopting free/open-source tools is framed as feminist praxis against gatekeeping in music tech
CLiC models a live-coding collective as horizontal and adhocratic — no owners, directors, or teachers
Latin American live coding communities — CLiC, TOPLAP MX, LiveCodeNet — are globally significant and often more Spanish-speaking than English-speaking in the TOPLAP ecosystem
The Mexican live-coding scene grew from a 2006–2014 institutional hub into a geographically dispersed network
Visual live-coding work is routinely under-credited relative to audio, and the gap tracks gender
Women have participated in live coding since its inception but face structural barriers; active advocacy and women-only spaces have been necessary to sustain diversity