Stack multi-part vocal harmonies, modulate extended acoustic chords, and sculpt dynamic swells.
Tap a slot to select or edit chord. Extended chords (maj9, add4, m11) create Lizzy McAlpine's distinctive bittersweet indie-folk voice leading.
Audition intimate lead vocal, lush 3rd above, resonant 5th below, and breathy whisper unison. Adjust stereo spread & dynamic gain.
Sculpt the emotional swell curve across the 4-bar phrase (controls filter cutoff, reverb wetness & stereo width).
The arranger loops four chord slots. Each slot lasts four beats, using sixty divided by the selected beats per minute, times four. At seventy eight beats per minute, a bar lasts about three point zero seven seven seconds, so a four bar loop lasts about twelve point three zero eight seconds. At seventy two, a bar is three point three three three seconds, and at eighty four it is two point eight five seven. At one hundred pixels per second, these bars have lengths about three hundred seven point seven, three hundred thirty three point three, and two hundred eighty five point seven. The timer advances a chord index modulo four. Changing tempo changes future timer scheduling, rather than a pre recorded soundtrack. The default dynamics array is point three, point five, point eight five, and point four. A higher value changes synthesized layer gain rather than chord duration. At four hundred pixels per unit the first three illustrated values become one hundred twenty, two hundred, and three hundred forty pixels. The high and low harmony gain combines its track volume with point five plus half the dynamic value. For track volume point seven, dynamics point three gives point four five five, and point eight five gives point six four seven five. That formula changes amplitude only; it does not guarantee a perceived emotion or reproduce a singer recording. The page has five mixer tracks: guitar, lead, upper harmony, lower harmony and whisper. The intimate preset mutes the upper and lower harmony tracks, leaving three unmuted tracks. The lyrical preset mutes only the lower track, leaving four. Default leaves five. At sixty pixels per unmuted track, the bars measure one hundred eighty, two hundred forty and three hundred. Solo can further restrict what is audible, so unmuted count is not identical to audible count. The JSON download captures tempo, chords, dynamics and all track settings. It is an arrangement specification, not exported audio or proof of professional mixing.
How does this arranger calculate voice leading, synthesis parameters, and dynamic gain across its multi-track layers?
This interactive arranger simulates a 4-bar acoustic progression paired with synthesized multi-track vocal harmonies. It translates chord voicings into discrete frequency targets for guitar plucks and dual-oscillator formant filters, while automating gain scaling based on dynamic swell curves.
The audio engine uses simplified Web Audio API dual oscillators and static formant bandpass filters ('a', 'o', 'u') rather than recorded acoustic samples or real-time pitch-tracking vocal formant models. Swell automation scales channel volume and strum velocity, not master reverb convolution or continuous dynamic filter sweeps.
Select Bar 3 (Gadd9) and adjust the Bar 3 swell slider in the Dynamic Section Arc Sculptor to 85% (0.85). The lead voice pitch is evaluated at A4 (440 Hz) with an effective gain factor of Vol * (0.6 + 0.85 * 0.4) = 0.85 * 0.94 ≈ 0.80. The lead sheet breakdown updates immediately to display 'Bar 3 (Gadd9): Lead Note: A4 | Harmony: D5 / G4 | Swell: 85%'.
Each chord in the internal dictionary defines specific MIDI-derived pitches for the acoustic plucked notes along with distinct voice targets: vocalLead, vocalHigh (3rd), and vocalLow (5th). Formant colors are approximated via dual bandpass BiquadFilter nodes (e.g., formant 'a' centered at 800 Hz and 1200 Hz with Q=4.0) fed by slightly detuned sawtooth and triangle oscillators.
Plucked acoustic guitar simulation relies on a triangle oscillator passed through an exponential-decay lowpass filter starting at 4.5 times the fundamental frequency and settling down toward 1.2 times fundamental, mimicking string attack dampening.
For each measure, the dynamic swell parameter (ranging from 0.10 to 1.0) scales the individual track volumes during step triggers. Lead vocal gain applies the formula vol * (0.6 + swell * 0.4), while harmony layers use vol * (0.5 + swell * 0.5), and whisper unison gain scales directly with swell.