I drew inspiration from Hans Zimmer’s work in Dune, with the industrial but desert sounds and the deep base. The sounds for the evil side of the track were also insipred by the Sardukar chant from the epilogue of each of the movies. The aesthetics of the hero soundtracks in games I played growing up.
The story being told is of a battle between evil and good a where good win after the beat drops (hopefully on time..). I have attached a video to the composition.
Looking at the video one more time, I realise that you can see the cursor bing visible in the middle of the screen. This is evidence of the p5js visualization shown as a video and there not being a p5 instance created during the performance.
The background visual is made with p5.js by taking muse mind monitors eeg readings of me playing chess and taking time-series data for each type of brain wave (alpha, delta, thetha) which are correlated to different types of emotions. The wave values, while in .csv had lost a lot of information and context where the color pf the particles is the aggregation of the wavetypes and the size is dictated by purely delta waves associated with high stress. The ring in the middle pulses on the heart-rate data from the headband.
The code code was used to generate the p5js video:
let brainData = [];
let headers = [];
let particles = [];
let currentIndex = 0;
let validData = [];
let heartBeatTimer = 0;
let lastHeartRate = 70; //default heart rate
class Particle {
constructor(x, y, delta, theta, alpha) {
this.pos = createVector(x, y);
this.vel = p5.Vector.random2D().mult(map(delta, -2, 2, 0.5, 3));
this.acc = createVector(0, 0);
this.r = map(delta, -2, 2, 50, 255);
this.g = map(theta, -2, 2, 50, 255);
this.b = map(alpha, -2, 2, 50, 255);
this.size = map(abs(delta), 0, 2, 2, 10);
this.lifespan = 255;
}
update() {
this.vel.add(this.acc);
this.pos.add(this.vel);
this.acc.mult(0);
this.vel.add(p5.Vector.random2D().mult(0.1));
this.lifespan -= 2;
}
applyForce(force) {
this.acc.add(force);
}
display() {
noStroke();
fill(this.r, this.g, this.b, this.lifespan);
ellipse(this.pos.x, this.pos.y, this.size, this.size);
}
isDead() {
return this.lifespan < 0;
}
}
function preload() {
brainData = loadStrings("mindMonitor_2025-02-26--20-59-08.csv");
}
function setup() {
createCanvas(windowWidth, windowHeight);
colorMode(RGB, 255, 255, 255, 255);
headers = brainData[0].split(",");
//filterinf dataset for valid rows
validData = brainData.slice(1).filter((row) => {
let cells = row.split(",");
return cells.some((cell) => !isNaN(parseFloat(cell)));
});
console.log("Total valid data rows:", validData.length);
}
function draw() {
//trailing effect for particles
background(0, 20);
if (validData.length === 0) {
console.error("No valid data found!");
noLoop();
return;
}
let currentLine = validData[currentIndex].split(",");
let hrIndex = headers.indexOf("Heart_Rate");
let heartRate = parseFloat(currentLine[hrIndex]);
if (!isNaN(heartRate) && heartRate > 0) {
lastHeartRate = heartRate;
}
let beatInterval = 60000 / lastHeartRate;
heartBeatTimer += deltaTime;
if (heartBeatTimer >= beatInterval) {
heartBeatTimer = 0;
let deltaIndex = headers.indexOf("Delta");
let thetaIndex = headers.indexOf("Theta");
let alphaIndex = headers.indexOf("Alpha");
let delta = parseFloat(currentLine[deltaIndex]) || 0;
let theta = parseFloat(currentLine[thetaIndex]) || 0;
let alpha = parseFloat(currentLine[alphaIndex]) || 0;
let particleCount = map(abs(delta), 0, 2, 5, 30);
for (let i = 0; i < particleCount; i++) {
let p = new Particle(
width / 2 + random(-100, 100),
height / 2 + random(-100, 100),
delta,
theta,
alpha
);
particles.push(p);
}
currentIndex = (currentIndex + 1) % validData.length;
}
push();
noFill();
stroke(255, 0, 0, 150);
strokeWeight(3);
let pulseSize = map(heartBeatTimer, 0, beatInterval, 100, 50);
ellipse(width / 2, height / 2, pulseSize, pulseSize);
pop();
for (let i = particles.length - 1; i >= 0; i--) {
particles[i].update();
particles[i].display();
if (particles[i].isDead()) {
particles.splice(i, 1);
}
}
}
function windowResized() {
resizeCanvas(windowWidth, windowHeight);
}
I encapsulated the tidal code so that i would only have to excute few lines during the performance which didn’t go the way I wanted:( , so in a way i feel like when it comes to live coding performance, having a few refactored functions work while working with code blocks lives gives a better setup for performing and debuggin on the fly.
Below is the tidal code:
-----------------hehe----
setcps (135/60/4)
-- harkonnen type beat
evilPad = slow 8
$ n (cat [
"0 3 7 10",
"3 6 10 13",
"5 8 12 15",
"7 14 17 21"
])
# s "sax"
# room 0.8
# size 0.9
# gain (range 0.7 0.9 $ slow 4 $ sine)
--shardukar type beat
evilBass = slow 8
$ sometimesBy 0.4 (rev)
$ n "0 5 2 7"
# s "sax:2"
# orbit 1
# room 0.7
# gain 0.75
# lpf 800
# shape 0.3
--geidi prime type beat
evilAtmosphere = slow 16
$ sometimes (|+ n 12)
$ n "14 ~ 19 ~"
# s "sax"
# gain 0.8
# room 0.9
# size 0.95
# orbit 2
--shardukar chant type pattern chpped
evilPercussion = slow 4
$ striate "<8 4>"
$ n (segment 16 $ range 6 8 $ slow 8 $ perlin)
# s "speechless"
# legato 2
# gain 1.2
# krush 4
--shardukar chant type pattern 2 chopped
evilVoice = slow 4
$ density "<1 1 2 4>/8"
$ striate "<2 4 8 16>/4"
$ n (segment 32 $ range 0 6 $ slow 8 $ sine)
# s "speech"
# legato (segment 8 $ range 2 3 $ slow 16 $ sine)
# gain (segment 32 $ range 0.8 1.2 $ slow 4 $ sine)
# pan (range 0.3 0.7 $ rand)
# crush 8
hush
evilRhythm = stack [
s "~ arp ~ ~" # room 0.5 # krush 9 # gain 1.2,
fast 2 $ s "moog2 moog2 moog2 moog3 moog:1 moog2" # room 0.7 # krush 5, fast 4 $ s "wobble8" # gain 0.8 # lpf 1200
]
d1 $ evilRhythm
hush
-- Build-up to drop
evilBuildUp = do {
d1 $ qtrigger $ filterWhen (>=0) $ seqP [
(0, 1, s "moog:14"), (1, 2, s "moog:18"),
(2, 3, s "moog:116"), (3, 4, s "moog:132")
] # room 0.3 # krush 9 # lpf (slow 4 (3000saw + 200)); d2 $ qtrigger $ filterWhen (>=0) $ seqP [ (0, 1, s "bass1:74"),
(1, 2, s "bass1:88"), (2, 3, s "bass1:916"),
(3, 4, s "bass1:932") ] # room 0.3 # lpf (slow 4 (1000saw + 100)) # speed (slow 4 (range 1 4 saw)) # gain 1.3;
d3 $ qtrigger $ filterWhen (>=0) $ seqP [
(0, 4, evilVoice # gain (slow 4 (range 0.8 1.5 saw)))
];
d4 $ qtrigger $ filterWhen (>=0) $ seqP [
(3, 4, s "crash:2*16" # gain (slow 1 (range 0.7 1.3 saw)) # room 0.9)
]
}
d1 $ silence
d2 $ silence
d3 silence
d4 silence
theDrop = do {
d1 silence;
d2 silence;
d3 silence;
d4 silence
}
heroBassDrum = s "[808bd:3(3,8), 808bd:4(5,16)]" # gain 1.3 # room 0.4 # shape 0.4 #krush 9
heroSnare = s "~ sd:2 ~ sd:2" # room 0.6 # gain 1.1 # squiz 0.8 #krush 6
heroHiHats = fast 2 $ s "hh*4" # gain (range 0.8 1.0 $ rand) # pan (range 0.2 0.8 $ slow 3 $ sine) #krush 5
heroTom = s "~ ~ ~ [~ lt:1 lt:2 lt:3*2]" # gain 1.0 # room 0.5 # speed 0.9 #krush 9
heroCymbal = s "[~ ~ ~ crash:4]" # gain 0.95 # room 0.7 # size 0.8
heroicFill = s "[feel:14, [~ ~ ~ feel:64], [~ ~ ~ ~ crash]]"
# gain 1.2
# room 0.7
# speed (range 1.5 0.9 $ saw)
# crush 8
dramaticEntrance = do {
d1 $ s "808bd:3 ~ ~ ~ ~ ~ ~ ~" # gain 1.4 # room 0.9 # size 0.9;
d2 $ s "~ ~ ~ crash:4" # gain 1.3 # room 0.9;
d3 silence;
d4 silence
}
heroPattern = do {
d1 $ stack [
heroBassDrum,
heroSnare,
heroHiHats
];
d2 $ stack [
heroTom,
heroCymbal
] # shape 0.3;
d3 $ s "sine:3" >| note (scale "mixolydian" ("0,4,7") + "c4")
# gain 0.65 # room 0.7;
d4 silence
}
heroExpanded = do {
d1 $ stack [
heroBassDrum # gain 1.4,
heroSnare # gain 1.2,
heroHiHats # gain 1.1
];
d2 $ stack [
heroTom,
heroCymbal # gain 1.1
] # shape 0.3;
d3 $ s "sine:3" >| note (scale "mixolydian" ("0,4,7") + "")
# gain 0.75 # room 0.7 # lpf 3000;
d4 $ s "[~ ~ ~ [feel:6 feel:7]]" # gain 0.9 # room 0.6
}
heroEpic = do {
d1 $ stack [
s "[808bd:3(3,8), 808bd:4(5,16)]" # gain 1.4 # room 0.4,
s "~ sd:2 ~ [sd:2 sd:4]" # room 0.6 # gain 1.2,
fast 2 $ s "hh4" # gain (range 0.9 1.1 $ rand) ]; d2 $ stack [ s "~ ~ mt:1 [~ lt:1 lt:2 lt:32]" # gain 1.1 # room 0.5,
s "[~ ~ ~ crash:4]" # gain 1.0 # room 0.7
] # shape 0.4;
d3 $ s "sine:3" >| note (scale "mixolydian" ("0,4,7,9") + "")
# gain 0.8 # room 0.8 # lpf 4000;
d4 $ s "feel:6(3,8)" # gain 0.9 # room 0.6 # speed 1.2
}
evilIntro = do {
d1 $ evilPad;
d2 silence;
d3 silence;
d4 silence
}
evilBuilds = do {
d1 $ evilPad;
d2 $ evilBass;
d3 $ evilAtmosphere;
d4 silence
}
evilIntensifies = do {
d1 $ evilPad;
d2 $ evilBass;
d3 $ evilAtmosphere;
d4 $ evilPercussion
}
d9 $ ccv "0 40 60 120" # ccn "0" # s "midi"
evilFullPower = do {
d1 $ evilPad # gain 0.6;
d3 $ evilVoice;
d4 $ evilPercussion # gain 1.3
}
--for evil full power
d6 $ s "~ arp ~ ~" # room 0.5 # krush 9 # gain 1.2
d7 $ fast 2 $ s "moog2 moog2 moog2 moog3 moog:1 moog2" # room 0.7 # krush 5 d8 $ fast 4 $ s "wobble8" # gain 0.8 # lpf 1200
evilIntro
evilBuilds
evilIntensifies
evilFullPower
evilBuildUp
theDrop
dramaticEntrance
heroPattern
heroExpanded
heroEpic
hushBelow is the hydra code:
s0.initVideo("/Users/janindu/Documents/liveCoding/p5_demo/p5_js_eegViz.mp4")
src(s0)
.blend(o1, () => cc[1])
.out()
src(o0)
.blend(src(o0).diff(s0).scale(.99),1.1)
.modulatePixelate(noise(2,0.8).pixelate(160,160),1024)
.out(o1)
src(o0)
.blend(
src(o0)
.diff(s0)
.scale(() => 0.99 + 0.1 * Math.sin(time * 0.1))
.rotate(() => cc[1] * 0.1),
() => 1.1 + cc[1] * 0.5
)
.modulatePixelate(
noise(() => 2 + cc[1] * 3, () => 0.01 + cc[1] * 0.05)
.pixelate(() => 16 - cc[1] * 8, () => 16 - cc[1] * 8),
() => 1024 * (1 - cc[1] * 0.5)
)
.modulate(
o0,
() => cc[1] * 0.3
)
.kaleid(() => Math.floor(cc[1] * 4) + 1)
.colorama(() => cc[1] * 0.1)
.out(o1)
//---
hush()
Im really happ something came out of the eeg pipieline but I honestly feel that that that time culd have been better spent relying on noise to generate visuals.
But nevertheless, I’m happy I could (pun intended..) leave my heart and soul on the screen.
Also this is how it looks like. from p5js:
