Tested on Processing 2.2.1 & 3.0a2 on OS X.
The code I've tweaked below may look familiar to some of you, it's what Imgur now uses as their loading animation. It was posted on OpenProcessing.org and I've been able to get it working in Processing, but the arcs are constantly wobbling around (relative movement within 1 pixel). I'm new to Processing and I don't see anything in the sketch that could be causing this, it runs in ProcessingJS without issue (though very high CPU utilization).
int num = 6;
float step, spacing, theta, angle, startPosition;
void setup() {
frameRate( 60 );
size( 60, 60 );
strokeWeight( 3 );
noFill();
stroke( 51, 51, 51 );
step = 11;
startPosition = -( PI / 2 );
}
void draw() {
background( 255, 255, 255, 0 );
translate( width / 2, height / 2 );
for ( int i = 0; i < num; i++ ) {
spacing = i * step;
angle = ( theta + ( ( PI / 4 / num ) * i ) ) % PI;
float arcEnd = map( sin( angle ), -1, 1, -TWO_PI, TWO_PI );
if ( angle <= ( PI / 2 ) ) {
arc( 0, 0, spacing, spacing, 0 + startPosition , arcEnd + startPosition );
}
else {
arc( 0, 0, spacing, spacing, TWO_PI - arcEnd + startPosition , TWO_PI + startPosition );
}
}
arc( 0, 0, 1, 1, 0, TWO_PI );
theta += .02;
}
If it helps, I'm trying to export this to an animated GIF. I tried doing this with ProcessingJS and jsgif, but hit some snags. I'm able to get it exported in Processing using gifAnimation just fine.
UPDATE
Looks like I'm going with hint( ENABLE_STROKE_PURE );, cleaned up with strokeCap( SQUARE ); within setup(). It doesn't look the same as the original but I do like the straight edges. Sometimes when you compromise, the result ends up even better than the "ideal" solution.
I see the problem on 2.2.1 for OS X, and calling hint(ENABLE_STROKE_PURE) in setup() fixes it for me. I couldn't find good documentation for this call, though; it's just something that gets mentioned here and there.
As for the root cause, if I absolutely had to speculate, I'd guess that Processing's Java renderer approximates a circular arc using a spline with a small number of control points. The control points are spaced out between the endpoints, so as the endpoints move, so do the bumps in the approximation. The approximation might be good enough for a single frame, but the animation makes the bumps obvious. Setting ENABLE_STROKE_PURE might increase the number of control points, or it might force Processing to use a more expensive circular arc primitive in the underlying graphics library it's built upon. Again, though, this is just a guess as to why a drawing environment might have a bug like the one you've seen. I haven't read Processing's source code to verify the guess.
Related
Some time ago, I coded a little fidgetable logo based on CSS transforms alone.
You can fiddle with it over https://document.paris/
The result feels nice, it feels natural to click/touch and drag to rotate the logo.
I remember banging my head against the walls until I found out that I could chain CSS transforms quite easily just by chaining them.
transform: matrix3d(currentMatrix) rotate3d(x, y, z, angle);
And most importantly to get the currentMatrix I would simply do m = $('#logobackground').css('transform'); with jQuery, the browser would magically return the computed matrix instead of the raw "css" which actually avoided me to deal with matrices or to infinitely stack rotate3D() properties.
So the hardest part was then to calculate the rotate3D arguments (x, y, z, angle) based on mouse inputs. In theory shouldn't have problems transposing this part to java so i'll just skip over it.
Now
I'm trying to do the exact same thing with Processing and there is two problems :
There is no rotate3D() in processing.
There is no browser to apply/chain transformations and return me the current matrix state automatically.
Here's the plan/implementation I'm working on :
I need a "currentMatrix" to apply every frame to the scene
PMatrix3D currentMatrix = new PMatrix3D();
In the setup() I set it to the "identity matrix" which from what I understand is equivalent to "no transformation".
// set currentMatrix to identity Matrix
currentMatrix.set(1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1);
Every frame I would calculate a transformation matrix and apply it to the currentMatrix.
Then I would apply this matrix to the scene.
// Apply Matrix to the currentMatrix
void mouseRotate() {
float diag = sqrt(pow(width,2)+pow(height,2));
float x = deltaX()/ diag * 10; // deltaX = difference between previous prevous MouseX and current mouseX)
float y = deltaY()/ diag * 10; // deltaY = same with Y axis
float angle = sqrt( pow(x, 2) + pow(y, 2) );
currentMatrix.apply( rotate3D(y,x,0,angle) );
}
// Apply Matrix to the scene
applyMatrix(currentMatrix);
PMatrix3D reference : https://processing.github.io/processing-javadocs/core/processing/core/PMatrix3D.html
ApplyMatrix() reference : https://processing.org/reference/applyMatrix_.html
All I need to do then is to implement the rotate3D css transform as a function which returns a transformation matrix.
Based on what I found on this page https://developer.mozilla.org/en-US/docs/Web/CSS/transform-function/rotate3d()
I implemented this first function :
PMatrix3D rotate3D(float x, float y, float z, float a) {
PMatrix3D rotationMatrix = new PMatrix3D();
rotationMatrix.set(
1+(1-cos(a))*(pow(x,2)-1), z*sin(a)+x*y*(1-cos(a)), -y*sin(a)+x*z*(1-cos(a)), 0,
-z*sin(a)+x*y*(1-cos(a)), 1+(1-cos(a))*(pow(y,2)-1), x*sin(a)+y*z*(1-cos(a)), 0,
y*sin(a)+x*z*(1-cos(a)), -x*sin(a)+y*z*(1-cos(a)), 1+(1-cos(a))*(pow(z,2)-1), 0,
0,0,0,1
);
return rotationMatrix;
}
and based on what I found on this page https://drafts.csswg.org/css-transforms-2/#Rotate3dDefined
I implemented this other function :
PMatrix3D rotate3Dbis(float getX, float getY, float getZ, float getA) {
float sc = sin(getA/2)*cos(getA/2);
float sq = pow(sin(getA/2),2);
float normalizer = sqrt( pow(getX,2) + pow(getY,2) + pow(getZ,2) );
float x = getX/normalizer;
float y = getY/normalizer;
float z = getZ/normalizer;
PMatrix3D rotationMatrix = new PMatrix3D();
rotationMatrix.set(
1-2*(pow(y,2)+pow(z,2))*sq, 2*(x*y*sq-z*sc), 2*(x*z*sq+y*sc), 0,
2*(x*y*sq+z*sc), 1-2*(pow(x,2)+pow(z,2))*sq, 2*(y*z*sq-x*sc), 0,
2*(x*z*sq-y*sc), 2*(y*z*sq+x*sc), 1-2*(pow(x,2)+pow(y,2)*sq), 0,
0, 0, 0, 1
);
return rotationMatrix;
}
When testing, they don't produce exactly the same result with the same inputs (although the differences are kind of "symmetric" which makes me think that they are kind of equivalent at least in some way ?) Also rotate3Dbis() has a tendency to produce NaN numbers, especially when i'm not moving the mouse (x & y = 0).
But most importantly, in the end it doesn't work. Instead of rotating, the drawing just zooms out progressively when I'm using rotate3D(), and rotate3Dbis() doesn't render correctly because of the NaNs.
The overall question :
I'm trying to get guidance from people who understand transformations Matrices and trying to narrow down where the issue is. Are my processing/java implementations of rotate3D() flawed ? Or would the issue come from somewhere else ? And are my rotate3D() and rotate3Dbis functions equivalent ?
You might get away with simply rotating on X and Y axis, as you already mentioned, using the previous and current mouse coordinates:
PVector cameraRotation = new PVector(0, 0);
void setup(){
size(900, 900, P3D);
rectMode(CENTER);
strokeWeight(9);
strokeJoin(MITER);
}
void draw(){
//update "camera" rotation
if (mousePressed){
cameraRotation.x += -float(mouseY-pmouseY);
cameraRotation.y += float(mouseX-pmouseX);
}
background(255);
translate(width * 0.5, height * 0.5, 0);
rotateX(radians(cameraRotation.x));
rotateY(radians(cameraRotation.y));
rect(0, 0, 300, 450);
}
The Document Paris example you've shared also uses easing. You can have a look at this minimal easing Processing example
Here's a version of the above with easing applied:
PVector cameraRotation = new PVector();
PVector cameraTargetRotation = new PVector();
float easing = 0.01;
void setup(){
size(900, 900, P3D);
rectMode(CENTER);
strokeWeight(9);
strokeJoin(MITER);
}
void draw(){
//update "camera" rotation
if (mousePressed){
cameraTargetRotation.x += -float(mouseY-pmouseY);
cameraTargetRotation.y += float(mouseX-pmouseX);
}
background(255);
translate(width * 0.5, height * 0.5, 0);
// ease rotation
rotateX(radians(cameraRotation.x -= (cameraRotation.x - cameraTargetRotation.x) * easing));
rotateY(radians(cameraRotation.y -= (cameraRotation.y - cameraTargetRotation.y) * easing));
fill(255);
rect(0, 0, 300, 450);
fill(0);
translate(0, 0, 3);
rect(0, 0, 300, 450);
}
Additionally there's a library called PeasyCam which can make this much simpler.
If you do want to implement your own version using PMatrix3D here are a couple of tips that could save you time:
When you instantiate PMatrix3D() it's the identity matrix. If you have transformations applied and you want to reset() to identity.
If you want to rotate a PMatrix3D() around and axis the rotate(float angleInRadians, float axisX, float axisY, float axisZ) override should help.
Additionally you could get away without PMatrix3D since resetMatrix() will reset the global transformation matrix and you can call rotate(float angleInRadians, float axisX, float axisY, float axisZ) directly.
Part of the answer is a fix added to the first rotate3D function.
I needed to normalize the x,y,z values to avoid the weird scaling.
I'm posting the current state of the code (i'm skipping a few parts for the sake of simplicity):
// Mouse movement since last fame on X axis
float deltaX() {
return (float)(mouseX-pmouseX);
}
// Mouse movement since last fame on Y axis
float deltaY() {
return (float)(mouseY-pmouseY);
}
// Convert user input into angle and amount to rotate to
void mouseRotate() {
double diag = Math.sqrt(Math.pow(width,2)+Math.pow(height,2));
double x = deltaX()/ diag * 50;
double y = -deltaY()/ diag * 50;
double angle = Math.sqrt( x*x + y*y );
currentMatrix.apply( rotate3D((float)y,(float)x,0,(float)angle) );
}
// Convert those values into a rotation matrix
PMatrix3D rotate3D(float getX, float getY, float getZ, float getA) {
float normalizer = sqrt( getX*getX + getY*getY + getZ*getZ );
float x = 0;
float y = 0;
float z = 0;
if (normalizer != 0) {
x = getX/normalizer;
y = getY/normalizer;
z = getZ/normalizer;
}
float x2 = pow(x,2);
float y2 = pow(y,2);
float z2 = 0;
float sina = sin(getA);
float f1cosa = 1-cos(getA);
PMatrix3D rotationMatrix = new PMatrix3D(
1+f1cosa*(x2-1), z*sina+x*y*f1cosa, -y*sina+x*z*f1cosa, 0,
-z*sina+x*y*f1cosa, 1+f1cosa*(y2-1), x*sina+y*z*f1cosa, 0,
y*sina+x*z*f1cosa, -x*sina+y*z*f1cosa, 1+f1cosa*(z2-1), 0,
0, 0, 0, 1
);
return rotationMatrix;
}
// Draw
draw() {
mouseRotate();
applyMatrix(currentMatrix);
object.render();
}
I thought that using this method would allow me to "stack" cumulative rotations relative to the screen and not relative to the object. But the result seems to always do the rotation relative to the object drawn.
I am not using a camera because I basically only want to rotate the object on itself. I'm actually a bit lost atm on what I should rotate and when to that the newly applied rotations are relative to the user, and the previously applied rotation are conserved.
So I am working with p5.js for class and I am very lost with it, as I dont understand very well. How do I animate this image to match with the sound? I tried frequency analysis but i dont know how to apply to the image. I wanted to animate it as i it was beating, like a heart, but according to the bpm sound i put in the sketch.
here is the sketch + image + sound
https://editor.p5js.org/FilipaRita/sketches/cUG6qNhIR
Actually finding the BMP for an entire piece of music would be a bit complicated (see this sound.stackexchange.com question), but if you just want to detect beats in real time I think you can probably hack something together that will work. Here is a visualization that I think will help you understand the data returned by fft.analyze():
const avgWindow = 20;
const threshold = 0.4;
let song;
let fft;
let beat;
let lastPeak;
function preload() {
song = loadSound("https://www.paulwheeler.us/files/metronome.wav");
}
function setup() {
createCanvas(400, 400);
fft = new p5.FFT();
song.loop();
beat = millis();
}
function draw() {
// Pulse white on the beat, then fade out with an inverse cube curve
background(map(1 / pow((millis() - beat) / 1000 + 1, 3), 1, 0, 255, 100));
drawSpectrumGraph(0, 0, width, height);
}
let i = 0;
// Graphing code adapted from https://jankozeluh.g6.cz/index.html by Jan Koželuh
function drawSpectrumGraph(left, top, w, h) {
let spectrum = fft.analyze();
stroke('limegreen');
fill('darkgreen');
strokeWeight(1);
beginShape();
vertex(left, top + h);
let peak = 0;
// compute a running average of values to avoid very
// localized energy from triggering a beat.
let runningAvg = 0;
for (let i = 0; i < spectrum.length; i++) {
vertex(
//left + map(i, 0, spectrum.length, 0, w),
// Distribute the spectrum values on a logarithmic scale
// We do this because as you go higher in the spectrum
// the same perceptible difference in tone requires a
// much larger chang in frequency.
left + map(log(i), 0, log(spectrum.length), 0, w),
// Spectrum values range from 0 to 255
top + map(spectrum[i], 0, 255, h, 0)
);
runningAvg += spectrum[i] / avgWindow;
if (i >= avgWindow) {
runningAvg -= spectrum[i] / avgWindow;
}
if (runningAvg > peak) {
peak = runningAvg;
}
}
// any time there is a sudden increase in peak energy, call that a beat
if (peak > lastPeak * (1 + threshold)) {
// print(`tick ${++i}`);
beat = millis();
}
lastPeak = peak;
vertex(left + w, top + h);
endShape(CLOSE);
// this is the range of frequencies covered by the FFT
let nyquist = 22050;
// get the centroid (value in hz)
let centroid = fft.getCentroid();
// the mean_freq_index calculation is for the display.
// centroid frequency / hz per bucket
let mean_freq_index = centroid / (nyquist / spectrum.length);
stroke('red');
// convert index to x value using a logarithmic x axis
let cx = map(log(mean_freq_index), 0, log(spectrum.length), 0, width);
line(cx, 0, cx, h);
}
<script src="https://cdnjs.cloudflare.com/ajax/libs/p5.js/1.3.1/p5.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/p5.js/1.3.1/addons/p5.sound.min.js"></script>
Hopefully this code with the comments helps you understand the data returned by fft.analyze() and you can use this as a starting point to achieve the effect you are looking for.
Disclaimer: I have experience with p5.js but I'm not an audio expert, so there could certainly be better ways to do this. Also while this approach works for this simple audio file there's a good chance it would fail horribly for actual music or real world environments.
If I were you then I would cheat and add some meta data that explicitly includes the timestamps of the beats. This would be a much simpler problem if you could shift the problem of beat detection to pre-processing. Maybe even do it by hand. Rather than trying to do it at runtime. The signal processing to do beat detection in an audio signal is non-trivial.
I have this code that basically reads each pixel of an image and redraws it with different shapes. All shapes will get faded in using a sin() wave.
Now I want to rotate every "Pixelshape" around its own axis (shapeMode(CENTER)) while they are faded in and the translate function gives me a headache in this complex way.
Here is the code so far:
void setup() {
size(1080, 1350);
shapeMode(CENTER);
img = loadImage("loremipsum.png");
…
}
void draw() {
background(123);
for (int gridX = 0; gridX < img.width; gridX++) {
for (int gridY = 0; gridY < img.height; gridY++) {
// grid position + tile size
float tileWidth = width / (float)img.width;
float tileHeight = height / (float)img.height;
float posX = tileWidth*gridX;
float posY = tileHeight*gridY;
// get current color
color c = img.pixels[gridY*img.width+gridX];
// greyscale conversion
int greyscale = round(red(c)*0.222+green(c)*0.707+blue(c)*0.071);
int gradientToIndex = round(map(greyscale, 0, 255, 0, shapeCount-1));
//FADEIN
float wave = map(sin(radians(frameCount*4)), -1, 1, 0, 2);
//translate(HEADACHE);
rotate(radians(wave));
shape(shapes[gradientToIndex], posX, posY, tileWidth * wave, tileHeight * wave);
}
}
I have tried many calculations but it just lets my sketch explode.
One that worked in another sketch where I tried basically the same but just in loop was (equivalent written):
translate(posX + tileWidth/2, posY + tileHeight/2);
I think I just don't get the matrix right? How can I translate them to its meant place?
Thank you very much #Rabbid76 – at first I just pasted in your idea and it went of crazy – then I added pushMatrix(); and popMatrix(); – turned out your translate(); code was in fact right!
Then I had to change the x and y location where every shape is drawn to 0,0,
And this is it! Now it works!
See the code here:
float wave = map(sin(radians(frameCount*4)), -1, 1, 0, 2);
pushMatrix();
translate(posX + tileWidth/2, posY + tileHeight/2);
rotate(radians(wave*180));
shape(shapes[gradientToIndex], 0, 0, tileWidth*wave , tileHeight*wave );
popMatrix();
PERFECT! Thank you so much!
rotate defines a rotation matrix and multiplies the current matrix by the rotation matrix. rotate therefore causes a rotation by (0, 0).
You have to center the rectangle around (0, 0), rotate it and move the rotated rectangle to the desired position with translate.
Since translate and rotate multiplies the current matrix by a new matrix, you must store and restore the matrix by pushMatrix() respectively popMatrix().
The center of a tile is (posX + tileWidth/2, posY + tileHeight/2):
pushMatrix();
translate(posX + tileWidth/2, posY + tileHeight/2);
rotate(radians(wave));
shape(shapes[gradientToIndex],
-tileWidth*wave/2, -tileHeight*wave/2,
tileWidth * wave, tileHeight * wave);
popMatrix();
I get different results when running this sample with Processing directly, and with Processing.js in a browser. Why?
I was happy about my result and wanted to share it on open Processing, but the rendering was totally different and I don't see why. Below is a minimal working example.
/* Program that rotates a triange and draws an ellipse when the third vertex is on top of the screen*/
float y = 3*height/2;
float x = 3*width/2;
float previous_1 = 0.0;
float previous_2 = 0.0;
float current;
float angle = 0.0;
void setup() {
size(1100, 500);
}
void draw() {
fill(0, 30);
// rotate triangle
angle = angle - 0.02;
translate(x, y);
rotate(angle);
// display triangle
triangle(-50, -50, -30, 30, -90, -60);
// detect whether third vertex is on top by comparing its 3 successive positions
current = screenY(-90, -60); // current position of the third vertex
if (previous_1 < previous_2 && previous_1 < current) {
// draw ellipse at the extrema position
fill(128, 9, 9);
ellipse(-90, -60, 7, 10);
}
// update the 2 previous positions of the third vertex
previous_2 = previous_1;
previous_1 = current;
}
In processing, the ellipse is drawn when a triangle vertex is on top, which is my goal.
In online sketching, the ellipse is drawn during the whole time :/
In order to get the same results online as you get by running Processing locally you will need to specify the rendering mode as 3d when calling size
For example:
void setup() {
size(1100, 500, P3D);
}
You will also need to specify the z coordinate in the call to screenY()
current = screenY(-90, -60, 0);
With these two changes you should get the same results online as you get running locally.
Online:
Triangle Ellipse Example
Local:
The problem lies in the screenY function. Print out the current variable in your processing sketch locally and online. In OpenProcessing, the variable current grows quickly above multiple thousands, while it stays between 0 and ~260 locally.
It seems like OpenProcessing has a bug inside this function.
To fix this however, I would recommend you to register differently when you drew a triangle at the top of the circle, for example by using your angle variable:
// Calculate angle and modulo it by 2 * PI
angle = (angle - 0.02) % (2 * PI);
// If the sketch has made a full revolution
if (previous_1 < previous_2 && previous_1 < angle) {
// draw ellipse at the extrema position
fill(128, 9, 9);
ellipse(-90, -60, 7, 10);
}
// update the 2 previous angles of the third vertex
previous_2 = previous_1;
previous_1 = angle;
However, because of how you draw the triangles, the ellipse is at an angle of about PI / 3. To fix this, one option would be to rotate the screen by angle + PI / 3 like so:
rotate(angle + PI / 3);
You might have to experiment with the angle offset a bit more to draw the ellipse perfectly at the top of the circle.
I have a p5.js sketch that draws radial gradients on the canvas in the browser window. They appear as they should, except when two or more overlap, when it looks like this: .
This is the class that is called to draw a radial gradient:
function Grey()
{
this.radius = int( random( 10, 200 ) );
this.x = random( 0 + this.radius, width - this.radius );
this.y = random( 0 + this.radius, height - this.radius );
this.display = function()
{
push();
for ( var i = 1; i <= this.radius; i++ )
{
var c = int( map( i, 1, this.radius, 0, 255 ) );
stroke( c );
ellipse( this.x, this.y, i, i );
}
pop();
};
}
edit: I have tried all available blending modes, neither was better than the default BLEND.
edit 2: code in p5.js editor
The overlapping ellipses are creating moiré patterns.
Moiré patterns are large-scale interference patterns that can be produced when an opaque ruled pattern with transparent gaps is overlaid on another similar pattern.
In this case we have circular lines that are closely spaced.
The essence of the moiré effect is the (mainly visual) perception of a distinctly different third pattern which is caused by inexact superimposition of two similar patterns.
This explains why in this question's code the pattern only appears where the sets of ellipses overlap. When overlap occurs we get a composite pattern that is similar but not exact to either of the individual patterns.
One easy way to lessen or nearly eliminate the moiré pattern behavior is to increase stroke weight.
In my experiments I can clearly see the pattern emerge with
strokeWeight(1);
and nearly disappear with:
strokeWeight(2);