/* webaudio.mjs - Copyright (C) 2022 Strudel contributors - see This program is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more details. You should have received a copy of the GNU Affero General Public License along with this program. If not, see . */ // import { Pattern, getFrequency, patternify2 } from '@strudel.cycles/core'; import * as strudel from '@strudel.cycles/core'; import { fromMidi, toMidi } from '@strudel.cycles/core'; import { loadBuffer } from './sampler.mjs'; const { Pattern } = strudel; import './vowel.mjs'; import workletsUrl from './worklets.mjs?url'; // export const getAudioContext = () => Tone.getContext().rawContext; let audioContext; export const getAudioContext = () => { if (!audioContext) { audioContext = new AudioContext(); } return audioContext; }; const getFilter = (type, frequency, Q) => { const filter = getAudioContext().createBiquadFilter(); filter.type = type; filter.frequency.value = frequency; filter.Q.value = Q; return filter; }; const getADSR = (attack, decay, sustain, release, velocity, begin, end) => { const gainNode = getAudioContext().createGain(); gainNode.gain.setValueAtTime(0, begin); gainNode.gain.linearRampToValueAtTime(velocity, begin + attack); // attack gainNode.gain.linearRampToValueAtTime(sustain * velocity, begin + attack + decay); // sustain start gainNode.gain.setValueAtTime(sustain * velocity, end); // sustain end gainNode.gain.linearRampToValueAtTime(0, end + release); // release // for some reason, using exponential ramping creates little cracklings return gainNode; }; const getOscillator = ({ s, freq, t, duration, release }) => { // make oscillator const o = getAudioContext().createOscillator(); o.type = s || 'triangle'; o.frequency.value = Number(freq); o.start(t); o.stop(t + duration + release); return o; }; const getSoundfontKey = (s) => { if (!globalThis.soundfontList) { // soundfont package not loaded return false; } if (globalThis.soundfontList?.instruments?.includes(s)) { return s; } // check if s is one of the soundfonts, which are loaded into globalThis, to avoid coupling both packages const nameIndex = globalThis.soundfontList?.instrumentNames?.indexOf(s); // convert number nameIndex (0-128) to 3 digit string (001-128) const name = nameIndex < 10 ? `00${nameIndex}` : nameIndex < 100 ? `0${nameIndex}` : nameIndex; if (nameIndex !== -1) { // TODO: indices of instrumentNames do not seem to match instruments return globalThis.soundfontList.instruments.find((instrument) => instrument.startsWith(name)); } return; }; const getSampleBufferSource = async (s, n, note) => { let transpose = 0; let midi; if (note !== undefined) { midi = typeof note === 'string' ? toMidi(note) : note; transpose = midi - 36; // C3 is middle C } const ac = getAudioContext(); // is sample from loaded samples(..) const samples = getLoadedSamples(); if (!samples) { throw new Error('no samples loaded'); } const bank = samples?.[s]; if (!bank) { throw new Error('sample not found:', s, 'try one of ' + Object.keys(samples)); } if (typeof bank !== 'object') { throw new Error('wrong format for sample bank:', s); } let sampleUrl; if (Array.isArray(bank)) { sampleUrl = bank[n % bank.length]; } else { if (!note) { throw new Error('no note(...) set for sound', s); } const midiDiff = (noteA) => toMidi(noteA) - midi; // object format will expect keys as notes const closest = Object.keys(bank) .filter((k) => !k.startsWith('_')) .reduce( (closest, key, j) => (!closest || Math.abs(midiDiff(key)) < Math.abs(midiDiff(closest)) ? key : closest), null, ); transpose = -midiDiff(closest); // semitones to repitch sampleUrl = bank[closest][n % bank[closest].length]; } const buffer = await loadBuffer(sampleUrl, ac); const bufferSource = ac.createBufferSource(); bufferSource.buffer = buffer; const playbackRate = 1.0 * Math.pow(2, transpose / 12); // bufferSource.playbackRate.value = Math.pow(2, transpose / 12); bufferSource.playbackRate.value = playbackRate; return bufferSource; }; const splitSN = (s, n) => { if (!s.includes(':')) { return [s, n]; } let [s2, n2] = s.split(':'); if (isNaN(Number(n2))) { return [s, n]; } return [s2, n2]; }; let workletsLoading; function loadWorklets() { if (workletsLoading) { return workletsLoading; } workletsLoading = getAudioContext().audioWorklet.addModule(workletsUrl); return workletsLoading; } function getWorklet(ac, processor, params) { const node = new AudioWorkletNode(ac, processor); Object.entries(params).forEach(([key, value]) => { node.parameters.get(key).value = value; }); return node; } try { loadWorklets(); } catch (err) { console.warn('could not load AudioWorklet effects coarse, crush and shape', err); } Pattern.prototype.out = function () { return this.onTrigger(async (t, hap, ct, cps) => { const hapDuration = hap.duration / cps; try { const ac = getAudioContext(); // calculate correct time (tone.js workaround) t = ac.currentTime + t - ct; // destructure value let { freq, s, sf, clip = 0, // if 1, samples will be cut off when the hap ends n = 0, note, gain = 1, cutoff, resonance = 1, hcutoff, hresonance = 1, bandf, bandq = 1, coarse, crush, shape, pan, attack = 0.001, decay = 0.05, sustain = 0.5, release = 0.001, speed = 1, // sample playback speed begin = 0, end = 1, vowel, } = hap.value; const { velocity = 1 } = hap.context; gain *= velocity; // legacy fix for velocity // the chain will hold all audio nodes that connect to each other const chain = []; if (typeof s === 'string') { [s, n] = splitSN(s, n); } if (typeof note === 'string') { [note, n] = splitSN(note, n); } if (!s || ['sine', 'square', 'triangle', 'sawtooth'].includes(s)) { // with synths, n and note are the same thing n = note || n || 36; if (typeof n === 'string') { n = toMidi(n); // e.g. c3 => 48 } // get frequency if (!freq && typeof n === 'number') { freq = fromMidi(n); // + 48); } // make oscillator const o = getOscillator({ t, s, freq, duration: hapDuration, release }); chain.push(o); // level down oscillators as they are really loud compared to samples i've tested const g = ac.createGain(); g.gain.value = 0.3; chain.push(g); // TODO: make adsr work with samples without pops // envelope const adsr = getADSR(attack, decay, sustain, release, 1, t, t + hapDuration); chain.push(adsr); } else { // load sample if (speed === 0) { // no playback return; } if (!s) { console.warn('no sample specified'); return; } const soundfont = getSoundfontKey(s); let bufferSource; try { if (soundfont) { // is soundfont bufferSource = await globalThis.getFontBufferSource(soundfont, note || n, ac); } else { // is sample from loaded samples(..) bufferSource = await getSampleBufferSource(s, n, note); } } catch (err) { console.warn(err); return; } // asny stuff above took too long? if (ac.currentTime > t) { console.warn('sample still loading:', s, n); return; } if (!bufferSource) { console.warn('no buffer source'); return; } bufferSource.playbackRate.value = Math.abs(speed) * bufferSource.playbackRate.value; // TODO: nudge, unit, cut, loop let duration = soundfont || clip ? hapDuration : bufferSource.buffer.duration; // let duration = bufferSource.buffer.duration; const offset = begin * duration; duration = ((end - begin) * duration) / Math.abs(speed); if (soundfont || clip) { bufferSource.start(t, offset); // duration does not work here for some reason } else { bufferSource.start(t, offset, duration); } chain.push(bufferSource); if (soundfont || clip) { const env = ac.createGain(); const releaseLength = 0.1; env.gain.value = 0.6; env.gain.setValueAtTime(env.gain.value, t + duration); env.gain.linearRampToValueAtTime(0, t + duration + releaseLength); // env.gain.linearRampToValueAtTime(0, t + duration + releaseLength); chain.push(env); bufferSource.stop(t + duration + releaseLength); } else { bufferSource.stop(t + duration); } } // master out const master = ac.createGain(); master.gain.value = gain; chain.push(master); // filters cutoff !== undefined && chain.push(getFilter('lowpass', cutoff, resonance)); hcutoff !== undefined && chain.push(getFilter('highpass', hcutoff, hresonance)); bandf !== undefined && chain.push(getFilter('bandpass', bandf, bandq)); vowel !== undefined && chain.push(ac.createVowelFilter(vowel)); coarse !== undefined && chain.push(getWorklet(ac, 'coarse-processor', { coarse })); crush !== undefined && chain.push(getWorklet(ac, 'crush-processor', { crush })); shape !== undefined && chain.push(getWorklet(ac, 'shape-processor', { shape })); // TODO delay / delaytime / delayfeedback // panning if (pan !== undefined) { const panner = ac.createStereoPanner(); panner.pan.value = 2 * pan - 1; chain.push(panner); } // master out /* const master = ac.createGain(); master.gain.value = 0.8 * gain; chain.push(master); */ chain.push(ac.destination); // connect chain elements together chain.slice(1).reduce((last, current) => last.connect(current), chain[0]); // disconnect all nodes when source node has ended: chain[0].onended = () => chain.forEach((n) => n.disconnect()); } catch (e) { console.warn('.out error:', e); } }); };