Modules
347 packages of audio work, each installed on its own: what decodes, measures, shapes, repairs, stretches and makes a sound. audio loads them by name.
No module by that name or for that. Try fewer words.
5 packagesbark, erb, gammatone, mel, octave
- barkBark-scale filter bank1.1.1
- erbERB-scale band centers and bandwidths (Glasberg & Moore) — pair with @audio/auditory-gammatone for actual filtering1.0.2
- gammatoneGammatone auditory filter (cochlear model)1.0.1
- melMel-frequency triangular filter bank1.0.1
- octaveIEC 61260 fractional-octave filter bank1.1.1
5 packagesdetect, beat-onset, tempo, track, onset
- detectFull beat detection pipeline: onset detection → tempo estimation → beat grid1.1.0
- beat-onsetSpectral flux onset detection1.0.4
- tempoTempo estimation via autocorrelation of onset detection function2.0.0
- trackDynamic programming beat tracker1.1.0
- onsetOnset detection — spectral-flux and energy-flux onset detection functions (ODFs) + adaptive peak picking1.0.3
25 packagesaudio-decode, aac, ac3, aiff, amr, ape, avi, caf, dsd, dts, eac3, flac, gsm, mod, mp3, mp4, mpc, opus, qoa, tta, vorbis, wav, wavpack, webm, wma
- audio-decodeDecode audio to AudioBuffer; unscoped alias of @audio/decode3.12.0
- aacDecode AAC/M4A (FAAD2 WASM) and ALAC/Apple Lossless (pure JS) to PCM samples1.6.0
- ac3Decode AC-3 (Dolby Digital) audio with liba52 WASM1.0.0
- aiffDecode AIFF/AIFF-C audio to PCM samples1.3.0
- amrDecode AMR-NB/WB audio via opencore-amr WASM1.3.1
- apeDecode Monkey's Audio (APE) with FFmpeg's ape decoder, WASM1.0.0
- aviDecode the audio stream of AVI files (PCM, MP3, AAC, µ-law, A-law) to PCM samples1.1.0
- cafDecode CAF (Core Audio Format) audio to PCM samples1.4.1
- dsdDecode DSD (DSF, DFF) audio to PCM samples1.0.0
- dtsDecode DTS (DTS Coherent Acoustics core) audio with libdca WASM1.0.0
- eac3Decode E-AC-3 (Dolby Digital Plus) and AC-3 with FFmpeg's eac3 decoder, WASM1.0.0
- flacDecode FLAC audio with libFLAC WASM2.0.0
- gsmDecode GSM 06.10 full rate audio (libgsm WASM)0.1.0
- modDecode tracker modules — MOD, XM, S3M, IT, MPTM and more — with libopenmpt WASM1.0.0
- mp3Decode MP3 audio with mpg123 WASM1.3.1
- mp4Decode the audio track of MP4, MOV, M4A, M4V and 3GP files (AAC, ALAC, MP3, FLAC, Opus, AMR, PCM) to PCM samples1.3.1
- mpcDecode Musepack (MPC, SV7 and SV8) audio with libmpcdec WASM1.0.0
- opusDecode Ogg Opus audio with local libopus WASM1.3.3
- qoaDecode QOA audio to PCM samples2.0.0
- ttaDecode TTA (True Audio) lossless audio to PCM samples2.0.0
- vorbisDecode Ogg Vorbis audio with libvorbis WASM1.3.4
- wavDecode WAV audio to PCM samples2.0.0
- wavpackDecode WavPack (.wv) audio with libwavpack WASM2.0.0
- webmDecode WebM / Matroska audio (Opus, Vorbis, AAC, MP3, FLAC, ALAC, PCM) to PCM samples1.6.1
- wmaDecode WMA audio (v1, v2, Pro, Lossless) with FFmpeg's decoders, WASM1.4.0
3 packagesanalyzer, notchbank, tracker
- analyzerFeedback candidate detection — spectral peaks with PNPR/PHPR howl criteria (Waterschoot & Moonen 2011)1.0.1
- notchbankAdaptive notch bank — pooled narrow IIR cuts, click-free coefficient morphing, deepen-on-redeploy1.2.0
- trackerHowl trajectory tracking — persistence + stability + narrowness gates; music discriminator1.0.1
20 packagesdebleed, debreath, declick, declip, decrackle, dehum, deplosive, dereverb, desilence, desqueak, detect, dewind, dewow, omlsa, repair, spectral, wiener, lpc, noise-estimate, vad
- debleedDe-bleed: a source's spill taken out of another microphone, given that source's own track. A Kalman filter cancels it through the room path it learns, a Wiener gain takes the residual it predicts0.3.0
- debreathDe-breath: breaths between phrases taken down, told by their length, level and the pauses either side; the room and the speech stay0.4.0
- declickDe-click: clicks found as outliers of the AR prediction error, each taken for a gap or a damped resonance by BIC, the resonance taken off, the gap rebuilt as its posterior mean under AR interpolation, everywhere or in given regions0.4.0
- declipDe-clip: each side's rail found where the sound was cut flat, or where lossy coding spread the cut into a band, or soft saturation's curve fitted blind by maximum likelihood and inverted, every clipped sample rebuilt three times, AR (Janssen) and sparse (A-SPADE) over 93 and 186 ms, the three blended by weights cross-validated region by region; unclipped sound untouched0.5.1
- decrackleDe-crackle: dense small impulses found as outliers of both the AR prediction error and the two-sided error, all of a window rebuilt at once by least-squares AR interpolation, looked for again until none is new, each event kept only where an impulse added to the sound explains it better than the sound's own excitation, the sound around them made robust by EM under impulsive noise0.4.0
- dehumMains hum, buzz and steady tone removal: the mains phase tracked, every line subtracted along it, then each tone held through the take (a whine, a test tone) found and subtracted; nothing found, no change0.7.1
- deplosiveDe-plosive: close-mic 'p', 'b' pops found as sudden aperiodic bursts under 80 Hz and taken out as a linear-phase low band, ~14 ms look-ahead0.3.0
- dereverbDe-reverb: late reverberation off a voice, one channel: weighted prediction error (WPE, Nakatani et al. 2010) fitted over the take, then the late power the past carries, scaled by the take's own decays, taken by a gain; a dry take or music passes untouched0.6.0
- desilenceDe-silence — VAD-driven silence handling for speech: remove, shorten (Smart Speed), split, trim0.2.1
- desqueakGuitar de-noise: string squeaks taken down where a finger slides on a wound string, the notes left as they ring; a pick's harsh attack softened, the amp's hiss and buzz taken down, on request0.3.0
- detectdenoise — content-aware auto-selector: finds the defect a recording carries, if any, and routes to its method0.7.1
- dewindDe-wind / de-rumble — per STFT bin, the wind between and under the harmonics taken, the harmonics kept0.4.0
- dewowWow & flutter correction — pitch-drift removal for tape/vinyl/cassette transfers0.3.0
- omlsaOM-LSA: Optimally Modified Log-Spectral Amplitude (Cohen & Berdugo 2001), with IMCRA noise (Cohen 2003) or a noise learned where it plays alone0.6.0
- repairSpectral repair: rebuild dropouts and damaged time×frequency regions by AR interpolation, a sinusoidal bridge or a similarity transplant, routed by length and content (iZotope RX class)0.4.0
- spectralSpectral subtraction (Boll 1979) with Berouti over-subtraction + spectral floor0.3.0
- wienerWiener filter / MMSE-LSA denoise with decision-directed a-priori SNR0.4.0
- lpcLinear predictive coding — Levinson-Durbin AR analysis, prediction, forward extrapolation, and least-squares gap interpolation1.2.0
- noise-estimateNoise PSD estimation — minimum statistics (Martin 2001), IMCRA (Cohen 2003), and quiet-segment profiling for spectral denoisers2.3.0
- vadVoice activity detection — frame-level speech decision (likelihood ratio over a minimum-statistics noise floor, anchored on voicing) plus per-bin speech-presence probability and decision-directed a-priori SNR2.0.1
dspThe signal processing the families are built on: filters, the FFT, windows, waveforms, weightings5 libraries
- digital-filterDigital filter design & processing: IIR, FIR, smoothing, adaptive, multirate2.4.3
- fourier-transformMinimalistic and efficient FFT implementation2.5.1
- window-functionWindow functions for signal processing and spectral analysis3.0.3
- periodic-functionPeriodic waveform functions — phase normalized to [0,1]2.1.2
- a-weightingA/B/C/D/Z/M frequency weighting functions for audio2.0.1
17 packagescompand, compressor, deesser, ducker, envelope, expander, fet, gate, leveler, limiter, multiband, opto, softclip, transient-shaper, unlimit, varimu, vca
- compandMulti-segment compander (SoX `compand`). Arbitrary piecewise-linear transfer0.1.7
- compressorFeed-forward soft-knee compressor, gain reduction smoothed in the log domain (Giannoulis, Massberg & Reiss 2012)0.3.1
- deesserDe-esser: the sibilance band's level over the voice body's drives a cut held within a range, read 5 ms ahead, on the band over 3.5 kHz (linear phase), a band, or broadband; the same at any recording level0.4.1
- duckerExternal-sidechain compressor: main signal's gain is reduced in proportion0.1.9
- envelopeBranching one-pole envelope follower. Returns a stateful per-sample detector0.1.4
- expanderDownward expander: reduces gain below threshold by (T-L)·(ratio-1) dB,0.2.2
- fetFET compressor model — 0.2 ms peak attack, high ratios, firm knee (1176 class)0.2.5
- gateNoise gate with hold-then-close logic and smoothed gain transitions0.1.7
- levelerGain riding / dialogue leveler — framewise smoothed loudness normalization (dynaudnorm / Vocal Rider class)0.2.1
- limiterLookahead brickwall limiter. A sliding-window maximum (monotonic deque) over0.2.1
- multibandMultiband compressor — Linkwitz-Riley split + per-band soft-knee compression, flat sum (SoX mcompand class)0.4.3
- optoOptical compressor model — RMS detector, wide knee, T4 program-dependent release (LA-2A class)0.2.5
- softclipStatic waveshaping — no envelope, no time state. Deterministic per-sample0.2.2
- transient-shaperTransient shaper — independently boosts/cuts attack and sustain portions0.1.6
- unlimitClassical de-limiter: transient-gated upward expansion restoring crest a brickwall limiter flattened0.1.2
- varimuVari-Mu compressor model — feedback detection, drive-dependent ratio, soft wide knee (Fairchild class)0.2.4
- vcaVCA compressor model — clean feed-forward, near-hard knee (dbx/SSL bus class)0.2.5
26 packagesautowah, bitcrusher, chorus, delay, distortion, exciter, flanger, freqshift, gain, graindelay, lofi, mixer, multitap, noiseshaper, phaser, pingpong, ringmod, rotary, sbr, slew, stutter, subbass, tapestop, tremolo, vibrato, wah
- autowahAuto-wah — envelope follower drives resonant bandpass cutoff1.1.5
- bitcrusherBitcrusher — sample-rate reduction + bit-depth quantization1.1.5
- chorusChorus — multiple detuned delay lines for ensemble thickening2.0.2
- delaySimple delay — mix of dry signal with delayed copy1.0.5
- distortionDistortion / saturation — non-linear waveshaping1.1.5
- exciterExciter — psychoacoustic harmonic synthesis for presence/air1.2.1
- flangerFlanger — modulated short delay with feedback2.0.2
- freqshiftFrequency shifter — single-sideband shift via Hilbert transform1.1.6
- gainGain — simple level adjustment in dB1.1.2
- graindelayGranular delay — per-grain pitch/time scatter (Ableton Grain Delay class)1.1.6
- lofiLo-fi character — wow/flutter, vinyl noise, crackle, bandwidth ceiling (RC-20 class)1.1.6
- mixerMixer — sums multiple buffers with gains1.0.2
- multitapMulti-tap delay — multiple delay taps at different times and gains1.1.6
- noiseshaperNoise shaping — error-feedback dithering for bit-depth reduction1.1.5
- phaserPhaser — cascade of swept first-order allpass filters creating moving notches/peaks1.1.6
- pingpongPing-pong delay — stereo: left and right delays alternate1.1.5
- ringmodRing modulator — multiplies signal by a carrier oscillator1.1.5
- rotaryRotary speaker (Leslie) simulation — split horn/drum rotors, Doppler + directivity AM0.1.2
- sbrSpectral band replication — extend HF from midband harmonics (aural exciter family; De-Slop bandwidth recovery)1.2.1
- slewRate-of-change limiter1.1.5
- stutterStutter / beat-repeat — sliced buffer retrigger (Ableton Beat Repeat class)1.1.6
- subbassPsychoacoustic bass enhancer — harmonic synthesis below cutoff (MaxxBass/RBass class)1.2.0
- tapestopTape-stop / spin-down — variable-rate playback with a decelerating read pointer0.1.2
- tremoloTremolo — amplitude modulation via LFO1.1.5
- vibratoVibrato — pitch modulation via modulated delay line2.0.2
- wahWah-wah — swept resonant bandpass filter1.1.5
14 packagesaac, aiff, alac, caf, flac, gsm, mp3, mp4, ogg, opus, qoa, wav, wavpack, webm
- aacEncode PCM audio samples to AAC (ADTS): WebCodecs, else the Fraunhofer FDK encoder (WASM)1.2.1
- aiffEncode PCM audio samples to AIFF format1.4.2
- alacEncode PCM audio to Apple Lossless (ALAC) — pure JS port of Apple's reference encoder1.0.2
- cafEncode PCM audio samples to CAF (Core Audio Format) LPCM1.3.2
- flacEncode PCM audio samples to FLAC format via WASM1.4.3
- gsmEncode PCM audio to GSM 06.10 full rate (libgsm WASM)0.1.0
- mp3Encode PCM audio samples to MP3 format via WASM2.0.1
- mp4Encode PCM audio samples to MP4/M4A (ISO BMFF) — pure JS muxer, remuxer and iTunes-tag writer1.2.1
- oggEncode PCM audio samples to OGG Vorbis format via WASM1.3.1
- opusEncode PCM audio samples to Ogg Opus (libopus WASM)1.3.2
- qoaEncode PCM audio samples to QOA (Quite OK Audio) format1.3.1
- wavEncode PCM audio samples to WAV format1.5.2
- wavpackEncode PCM audio samples to WavPack (.wv) with libwavpack WASM1.0.2
- webmEncode PCM audio samples to WebM (Opus, libopus WASM)1.3.1
10 packagesbaxandall, crossover, dynamic, fir, fit, graphic, highshelf, lowshelf, parametric, tilt
- baxandallBaxandall tone control — bass/treble shelving with reciprocal characteristics1.2.3
- crossoverN-way crossover network using Linkwitz-Riley filters1.0.2
- dynamicDynamic EQ — per-band peaking gain driven by band energy (Pro-Q3/soothe class, generalized de-esser)1.2.3
- firFIR convolution EQ — linear-phase, arbitrary magnitude response (frequency-sampling design; firequalizer / match-EQ class)1.0.2
- fitFit an N-band parametric EQ (RBJ peaking + shelves) to a target magnitude curve — greedy init + Levenberg–Marquardt refinement0.1.1
- graphicGraphic equalizer using ISO octave-band center frequencies1.2.4
- highshelfHigh-shelf filter — boost or cut above a corner frequency1.1.2
- lowshelfLow-shelf filter — boost or cut below a corner frequency1.1.2
- parametricN-band parametric EQ1.1.2
- tiltTilt EQ — see-saw around a pivot frequency1.2.3
16 packagesbiquad, filter-biquad, comb, dcblocker, derivative, diode-ladder, korg35, moog-ladder, oberheim, preemphasis, resonator, spectral-tilt, variable, speech-formant, speech-lpc, speech-vocoder
- biquadThe audio-side biquad — RBJ cookbook coefficients (Web Audio conventions) + stateful SOS kernel (process/step/cascade/filter)1.2.2
- filter-biquadAudio-facing biquad filters — highpass, lowpass, bandpass, notch, allpass (Hz/Q params, per-channel state)2.0.2
- combComb filter — feedforward (FIR) and feedback (IIR)1.1.3
- dcblockerDC blocking filter1.1.3
- derivativeDerivative / integral — FFmpeg aderivative/aintegral first-difference & running-sum pair1.0.2
- diode-ladderDiode ladder filter (Roland TB-303 / EMS VCS3 style)1.2.2
- korg35Korg MS-20 style filter (Korg35)1.2.1
- moog-ladderMoog 4-pole transistor ladder filter (-24 dB/oct lowpass with resonance; highpass, bandpass, 2-pole modes)1.2.0
- oberheimOberheim SEM state-variable filter (–12 dB/oct multimode)1.2.1
- preemphasisPre-emphasis / de-emphasis filter1.1.3
- resonatorConstant-peak-gain resonator (JOS two-zero form)1.2.1
- spectral-tiltSpectral tilt filter: apply a constant dB/octave slope to the spectrum1.2.1
- variableVariable-bandwidth filter with per-sample smoothed coefficient trajectory1.2.3
- speech-formantParallel formant filter bank1.0.1
- speech-lpcLinear Predictive Coding analysis/synthesis1.0.2
- speech-vocoderChannel vocoder1.1.1
7 packagescontrast, dr, lra, lufs, replaygain, sounds, truepeak
- contrastSpeech contrast — foreground/background RMS difference, WCAG 2.0 SC 1.4.7 (Audacity Contrast)1.0.4
- drDR value — TT/dr14 crest-factor dynamic range meter1.1.4
- lraLoudness range (LRA) — EBU Tech 3342 short-term distribution, P95−P10 after gating1.3.0
- lufsIntegrated loudness (LUFS) — ITU-R BS.1770-4 K-weighted gated measurement, EBU Tech 3341-verified1.2.0
- replaygainReplayGain 2.0 — track gain to the −18 LUFS reference1.2.1
- soundsSound labeling — level-threshold region detection (Audacity Label Sounds)1.0.3
- truepeakTrue peak (dBTP) — 4× sinc-oversampled inter-sample peak (BS.1770-5 Annex 2 methodology)1.1.7
7 packagesalign, correct, ir, latency, lossy, response, quality
- alignMulti-mic phase/time alignment — signed cross-correlation delay + polarity (Auto-Align class)1.0.2
- correctResponse correction design — measured response + target curve → FIR and/or parametric correction EQ (REW/Dirac/Acourate class)0.1.1
- irImpulse-response capture — exponential sine sweep deconvolution (Farina 2000), self-calibrated1.0.2
- latencyRound-trip latency — FFT cross-correlation peak between reference and loopback1.0.2
- lossyLossy-transcode detection — spectral-cutoff estimation + source classification (Spek / aucdtect class)0.1.1
- responseFrequency response from an impulse response — zero-padded FFT magnitude1.0.2
- qualityObjective audio quality metrics — SNR, segmental SNR, log-spectral distance, spectral similarity, noise-reduction ratio, per-partial modulation depth1.0.1
5 packagesparse, render, write, musicxml, subtitle
- parseStandard MIDI file parser — tracks, tempo map, note events1.0.3
- renderMIDI → PCM — GM-mapped synth render, no soundfont, no samples0.1.1
- writeStandard MIDI file writer — note events → SMF (mir-transcribe output target)1.0.3
- musicxmlMusicXML 4.0 writer/parser and ABC writer — note events ↔ notation0.2.0
- subtitleTimed-text subtitle formats — SRT, WebVTT, LRC (+ enhanced word-level), TTML: parse and write0.1.1
16 packageschord, chordino, chroma, coversong, downbeat, drums, fingerprint, key, melody, multif0, nnls-chroma, similarity, structure, tempogram, tonnetz, transcribe
- chord24 binary chord templates: C, C#, … B (major) then C, C#, … B (minor)1.0.2
- chordinoChordino (Mauch 2010): chord sequence from NNLS chroma, as the reference Vamp plugin1.0.0
- chromaChroma (pitch-class profile) feature — a 12-D vector where each bin holds1.1.3
- coversongCover song identification — recognize the same composition across performances1.1.5
- downbeatDownbeat estimation — locate the bar-level "1" within a beat grid1.1.5
- drumsDrum transcription — kick/snare/hihat onset detection and classification1.1.6
- fingerprintLandmark fingerprinting — spectral peak constellation hashes + offset-vote matching (Wang 2003 class)1.1.4
- keyKrumhansl–Kessler major and minor key profiles1.0.2
- melodyMelody contour — frame-level F0 track via YIN with voicing (MIREX melody shape)1.1.4
- multif0Multiple F0 estimation — all simultaneous pitches per frame (polyphonic)1.1.5
- nnls-chromaNNLS Chroma (Mauch & Dixon 2010): treble and bass chromagram with tuning, as the reference Vamp plugin1.0.0
- similarityAudio similarity — distance metric between recordings1.1.5
- structureStructural segmentation — Foote checkerboard novelty over MFCC self-similarity (Foote 2000)1.1.4
- tempogramTempogram — local tempo over time via onset-envelope autocorrelation1.1.5
- tonnetzTonal centroid (Tonnetz) — 6-dim interval-circle projection of chroma (Harte 2006)1.1.3
- transcribePolyphonic transcription — audio → MIDI note events (onset, offset, pitch, velocity)1.1.5
8 packagesamdf, autocorrelation, cepstrum, hps, mcleod, pyin, swipe, yin
- amdfAMDF (Average Magnitude Difference Function) pitch detection1.0.2
- autocorrelationNormalized autocorrelation pitch detection (baseline)1.0.2
- cepstrumCepstrum pitch detection (Noll, 1967)1.0.2
- hpsHarmonic Product Spectrum (Schroeder, 1968)1.0.2
- mcleodMcLeod Pitch Method (McLeod & Wyvill, 2005)1.0.2
- pyinpYIN — probabilistic YIN (Mauch & Dixon, 2014)2.1.0
- swipeSWIPE′ — Sawtooth Waveform Inspired Pitch Estimator with prime harmonics1.0.2
- yinYIN pitch detection (de Cheveigné & Kawahara, 2002)1.0.6
platformThe Web Audio API and its AudioBuffer, the same in Node as in a browser2 libraries
- resampleSample-rate conversion — umbrella for @audio/resample-* atoms (sinc, linear, polyphase)1.2.4
3 packageslinear, polyphase, sinc
- linearLinear-interpolation resampling — cheap, real-time (no anti-alias; use resample-sinc for downsampling)1.0.3
- polyphasePolyphase FIR resampling — streaming-friendly rational rate conversion (48↔16/24 kHz voice-agent path)1.1.0
- sincWindowed-sinc (Lanczos, 32-tap) resampling with anti-alias kernel widening on downsample1.3.0
7 packagesconvolution, dattorro, fdn, freeverb, schroeder, shimmer, spring
- convolutionConvolution reverb — impulse-response convolution (rooms, plates, cabinets)1.1.0
- dattorroDattorro plate reverb — input diffusers + figure-eight tank (JAES 1997 topology and taps)1.1.3
- fdnFDN reverb — 8 mutually-prime delays, O(N) Householder feedback, per-line damping (Jot topology)1.1.3
- freeverbFreeverb — Schroeder-Moorer 8-comb + 4-allpass reverb, Jezar public-domain tuning, stereo spread1.0.5
- schroederSchroeder reverb — 4 parallel feedback combs with LP damping + 2 series allpass diffusers1.1.3
- shimmerShimmer reverb — octave-up pitch-shifted feedback tail (Valhalla/Eno-Lanois class)1.1.4
- springSpring reverb — dispersive allpass-chain loop (Parker-Välimäki class simplified model)1.1.3
5 packagesmultiband, tape, transistor, tube, waveshaper
- multibandMultiband saturation — LR split + per-band tube drive (Saturn class)1.1.4
- tapeTape saturation — tanh transfer + playback HF rolloff (simplified machine model)1.1.3
- transistorTransistor/console saturation — odd-symmetric cubic, 3rd-harmonic dominant1.1.3
- tubeTube saturation — asymmetric biased tanh, even-harmonic signature1.0.5
- waveshaperWaveshaper — normalized tanh or custom curve, oversampled1.1.3
15 packagesdelay, formant, granular, hpss, hybrid, lpc, ola, paulstretch, psola, pvoc, pvoc-lock, sample, sms, transient, wsola
- delayDelay-line (harmonizer) pitch shift — dual crossfading taps sweeping a modulated delay window1.1.0
- formantFormant-preserving pitch shift via cepstral envelope re-imposition1.1.5
- granularGranular pitch shift — direct grain-read synthesis for signature textural sound1.1.0
- hpssHarmonic/Percussive Source Separation pitch shift — shifts harmonics, preserves transients1.1.0
- hybridHybrid pitch shift — runs phase-lock + WSOLA in parallel, crossfades by transient confidence1.1.0
- lpcLPC source-filter pitch shift — repitched excitation residual through the unmodified all-pole vocal-tract filter1.1.0
- olaOLA (Overlap-Add) pitch shift — simplest stretch+resample baseline1.1.0
- paulstretchPaulstretch-style randomized-phase pitch shift for textural blur1.1.5
- psolaPSOLA (Pitch-Synchronous Overlap-Add) pitch shift — pitch-synchronous grains for monophonic voice1.1.0
- pvocCanonical phase vocoder (Bernsee/SMB) pitch shift1.1.5
- pvoc-lockPeak-locked phase vocoder (Laroche-Dolson) pitch shift1.0.3
- sampleSampler-style pitch shift — sinc-interpolated fractional-stride resample (no time preservation)1.1.0
- smsSpectral Modeling Synthesis (Serra/Smith) sinusoidal+residual pitch shift1.1.0
- transientTransient-aware phase vocoder pitch shift with attack preservation1.1.0
- wsolaWSOLA (Waveform-Similarity Overlap-Add) pitch shift — similarity-search OLA variant1.1.0
3 packagesresidual, synth, track
- residualResidual noise layer — spectral subtraction of tracked partials with input phase (Serra SMS)1.0.2
- synthAdditive resynthesis from partial trajectories — interpolated freq/amp, integrated phase1.0.2
- trackSinusoidal analysis — STFT peaks (parabolic) + nearest-frequency partial tracking (McAulay-Quatieri 1986)1.0.2
12 packagesambisonic, autopan, binaural, channelsplit, crossfeed, delay, haas, microshift, midside, panner, surround, widener
- ambisonicAmbisonics — encode/rotate/decode, 1st-3rd order (HOA), ACN/SN3D (AmbiX), FuMa conversion0.1.0
- autopanAuto-panner — LFO-driven stereo panning. Constant-power cos/sin law1.1.3
- binauralStructural binaural panner — Brown-Duda spherical-head ITD + head-shadow ILD (no HRIR data)0.1.1
- channelsplitChannel split/combine — multichannel ↔ mono planes (FFmpeg channelsplit)1.0.1
- crossfeedHeadphone crossfeed filter1.1.1
- delayPer-channel delay in ms/samples (FFmpeg adelay)1.0.1
- haasHaas effect — delays one channel by 1–35 ms to create phantom stereo image1.1.3
- microshiftMicro-pitch stereo widener — ±cents detuned copies (H3000 MicroShift class)1.1.3
- midsideMid/side encode-decode — L/R ↔ M/S matrix with width control1.1.3
- pannerStereo panner — constant-power panning1.1.3
- surroundStereo → 5.1 matrix upmix — C/LFE/Ls/Rs derivation (documented simple matrix)1.1.3
- widenerStereo widener — mid/side processing to adjust stereo width1.1.3
17 packagescentroid, contrast, cqt, crest, edit, flatness, flux, freeze, harmonics, ltas, mfcc, pvoc, rolloff, slope, spread, target, zcr
- centroidSpectral centroid — magnitude-weighted mean frequency (Hz). Peeters 2004 §6.1.1.0.3
- contrastSpectral contrast — per-octave peak/valley difference (Jiang 2002, librosa parity)1.1.4
- cqtConstant-Q transform — log-frequency spectrum, Q-scaled windows (Brown 1991)1.0.2
- crestSpectral crest — peak-to-mean ratio of the POWER spectrum (≥1; high = peaky/tonal).1.0.3
- editSpectral editing — gain on time×frequency regions via COLA-normalized STFT (Audacity/afftfilt class)1.0.3
- flatnessSpectral flatness — geometric/arithmetic mean ratio of the POWER spectrum, 0 (tonal) … 1 (noise).1.0.3
- fluxSpectral flux — frame-to-frame magnitude change. Euclidean by default (FFmpeg aspectralstats);1.1.5
- freezeSpectral freeze — sustain a spectral frame with phase dispersion (Spectral Time class)1.0.2
- harmonicsHarmonic descriptors — inharmonicity, tristimulus, odd/even ratio from f0 (essentia parity)1.0.3
- ltasLTAS — long-term average spectrum via Welch's method (adaptive-EQ / match-EQ substrate)1.1.5
- mfccMFCC — mel-frequency cepstral coefficients (FFT → mel triangles → log → DCT-II)1.0.3
- pvocPhase-vocoder engine — peak-gated and peak-locked spectral bin scatter (Bernsee, Laroche-Dolson), phase locking, peak finding; powers @audio/shift-pvoc* and @audio/stretch-pvoc*1.1.1
- rolloffSpectral rolloff — frequency below which \`p\` of total spectral energy lies (default 85%).1.1.5
- slopeSpectral slope — least-squares linear regression of magnitude over frequency. Peeters 2004 §6.6.1.1.5
- spreadSpectral spread — magnitude-weighted standard deviation around the centroid (Hz). Peeters 2004 §6.1.1.1.5
- targetTarget-curve library — per-content-type LTAS targets (speech/music/pink/voice-music) + deviation math for adaptive/match EQ1.1.0
- zcrZero-crossing rate — per-frame sign-change fraction (librosa zero_crossing_rate class)1.0.3
10 packageshybrid, paulstretch, pghi, psola, pvoc, pvoc-lock, pvsola, sms, transient, wsola
- hybridHybrid harmonic/percussive time stretch (Driedger) — HPSS split, phase-locked vocoder + OLA per layer1.2.3
- paulstretchPaulstretch — Paul Nasca's extreme stretch (8×+) for ambient/drone material1.2.2
- pghiPhase Gradient Heap Integration vocoder (Průša & Holighaus "Phase Vocoder Done Right") — no peak picking, no transient heuristics1.3.0
- psolaPitch-Synchronous Overlap-Add — time-domain stretch for speech and monophonic instruments1.2.3
- pvocPhase vocoder time stretching — educational baseline without phase locking1.2.3
- pvoc-lockPhase-locked vocoder (Laroche & Dolson 1999) — eliminates phasiness of plain pvoc1.3.3
- pvsolaPVSOLA time stretching: a phase-locked vocoder reset to the input's own waveform where it fits; speech slowed without flanging or phasiness1.0.0
- smsSinusoidal Modeling Synthesis time stretching — harmonic/tonal material1.2.2
- transientTransient-aware phase-locked vocoder (Röbel 2003) — preserves attacks on drums and plucks1.2.3
- wsolaWSOLA time stretching — time-domain: speech and solo voices keep their pulse shape and attacks2.0.0
20 packageschirp, drum, dtmf, dx7, envelope, fm, lfo, modal, modal-fit, noise, osc, pluck, poly, rhythm, risset, sfx, tonewheel, tonewheel-fit, voice, wavetable
- chirpChirp — linear/exponential frequency sweep (ESS measurement signal, Farina 2000)1.1.3
- drumDrum synthesis — membrane kick, metal cymbal, noise snare (Tone.js class)2.0.2
- dtmfDTMF dial tones — ITU-T Q.23 row/column pairs per digit1.0.1
- dx7Yamaha DX7 FM engine: a bit-exact JavaScript port of the Music Synthesizer for Android core (6 operators, 32 algorithms, DX7 envelopes, LFO, pitch EG), with DX7 sysex loading0.1.0
- envelopeADSR envelope generator — gain contour for amplitude/filter modulation1.1.3
- fmFM synthesis — Chowning phase modulation, single-op or serial op stack, bell/epiano presets0.2.1
- lfoLFO — low-frequency control signal, bipolar/unipolar1.0.1
- modalModal synthesis — two-pole resonator bank (string/bar/membrane/plate/tube), impulse or noise excited0.2.2
- modal-fitModal snapshot from audio: one strike → synth-modal's { freq, amp, modes: [{ ratio, gain, t60 }] } (subband matrix pencil, least-squares amplitudes)0.1.1
- noiseColors of noise — white, pink (Kellet), brown, blue, violet; seeded, slope-verified2.0.2
- oscOscillator — band-limited sine/square/sawtooth/triangle with detune1.1.0
- pluckKarplus-Strong plucked string synthesis1.1.4
- polyPolyphonic voice allocator — wrap any voice with N-voice polyphony1.2.1
- rhythmRhythm / click track — BPM grid with accented downbeats1.1.3
- rissetRisset drum — inharmonic partial set with exponential decay1.1.3
- sfxGame/UI sound effects generator — ZZFX / ChipTone class parameterized synthesis1.1.4
- tonewheelTonewheel organ (Hammond B-3 upper manual): 91 geared tonewheels, nine drawbars, foldback, single-trigger percussion, key click, vibrato/chorus scanner0.1.0
- tonewheel-fitTonewheel registration from audio: sustained organ note or chord → nine drawbars 0..8, percussion guess, score (NNLS against the tonewheel model)0.1.0
- voiceSynth voice — osc × ADSR through envelope-scaled lowpass (Tone.js Synth class)1.2.0
- wavetableMorphing wavetable oscillator — scan a bank of single-cycle tables (WaveEdit class)1.0.1
3 packagescurve, midi, snap
- curveRetune a voice along a pitch curve, its own glottal cycles re-spaced (TD-PSOLA): formants and consonants kept, streaming (Melodyne class)0.4.0
- midiNote-target pitch correction — align melody to reference MIDI/notes (Melodyne class)1.0.2
- snapPitch correction to a scale: each note's centre moved onto its note, its slow drift flattened, the voice's own cycles re-spaced along one pitch curve (Auto-Tune / Melodyne class)2.0.2
uiA sound shown and edited on a page: a waveform, a spectrogram, a typeface, a text field4 libraries
- waveareaEdit audio in textarea1.1.0
- gl-waveformWebGL2 renderer for numeric signals, time series and audio: exact peaks, RMS band, sample-accurate line5.3.0
- gl-spectrogramWebGL2 spectrogram renderer: FFT on the GPU, reassigned to one-pixel lines, sharp at every zoom2.4.0
- wavefontTypeface for visualizing waveform data3.8.2
utilSmall helpers for audio data: its format told from its bytes, its samples from one layout to another2 libraries
3 packagesglottis, tract, voder
7 packagesa, b, c, d, itu468, k, riaa
- aIEC 61672-1:2013 analog prototype frequencies1.2.3
- bIEC 61672-1:2013 / ANSI S1.42 B-weighting filter — moderate-loudness frequency response1.1.3
- cIEC 61672-1:2013 C-weighting: H(s) = K·s² / ((s+w1)²(s+w4)²)1.2.3
- dIEC 537 D-weighting filter — aircraft-noise frequency response (resonant hump near 4 kHz)1.0.3
- itu468ITU-R BS.468-4 noise weighting, exact matched-z realization: peaked at1.2.3
- kITU-R BS.1770-4 K-weighting: high-shelf + RLB highpass1.2.1
- riaaRIAA playback (de-emphasis): H(s) = (1 + s*T2) / ((1 + s*T1) * (1 + s*T3))1.1.3