I want to measure the performance of browsers from my JavaScript when they render the HTML5 Canvas.
To begin, I drew a simple 2D canvas and logged the time right before and after it was done:
HTML
<canvas id="myCanvas" width="200" height="100"></canvas>
JS
var start = performance.now();
var c = document.getElementById("myCanvas");
var ctx = c.getContext("2d");
ctx.moveTo(0, 0);
ctx.lineTo(200, 100);
ctx.stroke();
var end =performance.now();
alert(end-start); //alert for testing only
This returned a very small value as predicted.
To mesure the performance I should render something more complex, So i did a code that renders a 3D image and animates it.
But now It's tricky to calculate time before and after since the animation is on a constant loop. I tried logging the time for the first animation before it continued to loop, but still the time difference is very small.
I want to see if I am in the right direction or if there are better practices. My main aim is to check differences in performance between as many browsers as possible, using javascript, and I thought this was better logged while doing a complex task such as renderign a 3D canvas.
Related
I am trying to verify that this happens no matter what, and there's no way to bypass it. It seems pretty silly to me that createjs uses this architecture. But I noticed that creating a stage from an existing canvas element removes all shapes, writings, etc from the canvas? Code is below:
html:
<canvas id="canvas" width="800" height="400"></canvas>
js:
var canv = document.getElementById("canvas");
var stage = new createjs.Stage(canv);
var ctx = canv.getContext('2d');
ctx.beginPath();
ctx.arc(50,50,10,0,2*Math.PI);
ctx.stroke();
createjs.Ticker.addEventListener("tick", tick);//circle created is overwritten here!!!?! Why createjs!?
//stage.update();
createjs.MotionGuidePlugin.install();
var shape1 = new createjs.Shape();
shape1.graphics.f("#000000").dc(0,0,10);
var path1 = new createjs.Shape();
path1.graphics.beginStroke("#ff0000").mt(0,0).qt(50,100,100,0);
stage.addChild(path1, shape1);
createjs.Tween.get(shape1).to({guide:{ path:[0,0, 50,100, 100,0] }},2000, createjs.Ease.cubicInOut);
function tick(event) {
stage.update();
}
Is this something that cannot be bypassed? It seems silly to me that createjs wouldn't just actually use the existing element unerased. If not, what is the point in passing in an element in the first place, why doesn't it just create a canvas element, and give you the ID? Anyways, if this is how it's going to be, unfortunately, I am going to have to go somewhere else. Sad, because createjs seemed pretty useful.
CreateJS uses a retained graphics mode, that is, it stores the state and redraws it each time. This is because the canvas is basically a big Bitmap with drawing commands – clearing the stage is the only way to remove the previous state.
But good news! There are lots of ways to get around these limitations if you want to blend CreateJS content with other content, or even make additive drawing effects.
The first is easy, which is setting autoClear. This will prevent the clear, and just draw the new contents over the old one.
stage.autoClear = false;
The second is a bit tougher, but great for instances where you want to mix CreateJS content with other libraries or effects. You can basically use the other canvas as the source to a Bitmap you include in CreateJS:
// Create a child for CreateJS referencing the other canvas
var bmp = new createjs.Bitmap(otherCanvas);
// Add it at the bottom (or top or wherever) in your new CreateJS canvas
stage.addChildAt(bmp, 0);
This is a great approach because it lets you put your content wherever you want, edit it separately, etc.
If you have a different case this doesn't cover, let me know and I can try to make a recommendation!
Cheers.
I'm looking for a way to render graphics onto an HTML5 canvas using JavaScript, but I want to only render said graphics if they're inside a pre-defined mask.
I'm creating a GUI framework that can be used to easily and quickly create GUIs on an HTML5 canvas. I think that something that would be really nice to have is a way to render graphics inside an element, and make the element auto-crop the graphics so that they always stay inside of it. For example, I can make a rectangular element and animate a circular pulse inside of it, and as the circle extends past the outside of the element, those parts of he circle should just not render to keep it looking smooth and sharp. This is similar to what CSS does with overflow: hidden;
Now, I know that one option is to use a mask-like feature. For example, P5.js has mask(). However, this is very very slow. Masking a single element a single time using P5.js significantly reduces framerate, and I want to be doing this potentially hundreds of times per frame without frame drops. I know that CSS does this incredibly efficiently (from my own experience working with it), but I can't seem to think of any way to make it efficient on a canvas element.
I could do it pretty simply if it was just a rectangle, but I want to do this for any shape. For example, a circle, a star, a rectangle with rounded edges, or really any polygon at all.
How can this be done? I thought of potentially rendering to an off screen canvas (which is shrunken to the size of the element in question), then render the element onto that screen using one color (let's say the background color will be white, and the shape will be black), then rendering the image we want masked onto another off screen canvas that's the same width as our other OSC, then looping through one of their image data arrays and mapping one to the other based on whether said pixel is white or black on the mask canvas.
But........ I can't help but think that that's going to be incredibly slow for the computer to process. I assume that CSS somehow leverages the GPU to do this type of computation incredibly efficiently and that's why they get such an increase in performance. Is it possible for me to do the same or am I just dreaming?
Okay, so I have found two different means of doing this (huge thank you to #Kaiido). One method is to use ctx.clip() while one works with CanvasPattern.
This snippet shows both means in action:
<canvas id = "c" width = "400" height = "400"></canvas>
<canvas id = "c2" width = "400" height = "400"></canvas>
<script>
var canvas = document.getElementById("c");
var ctx = canvas.getContext("2d");
ctx.fillStyle = "yellow";
ctx.fillRect(0,0,400,400);
ctx.beginPath();
ctx.arc(200,200,100,0,6);
ctx.clip();
ctx.beginPath();// This clears our previous arc from the path so that it doesn't render in when we `fill()`
ctx.fillStyle = "rgb(255,0,0)";
for(var i = 0;i < 20;i++){
for(var j = 0;j < 40;j++){
ctx.rect(i * 20 + j % 2 * 10,j * 10,10,10);
}
}
ctx.fill();
</script>
<script>
var canvas2 = document.getElementById("c2");
var ctx2 = canvas2.getContext("2d");
ctx2.fillStyle = "orange";
ctx2.fillRect(0,0,400,400);
var osc = new OffscreenCanvas(400,400);
var oscctx = osc.getContext("2d");
oscctx.fillStyle = "rgb(255,0,0)";
for(var i = 0;i < 20;i++){
for(var j = 0;j < 40;j++){
oscctx.rect(i * 20 + j % 2 * 10,j * 10,10,10);
}
}
oscctx.fill();
var pattern = ctx2.createPattern(osc,"no-repeat");
ctx2.fillStyle = pattern;
ctx2.arc(200,200,100,0,6);
ctx2.fill();
</script>
Which one is more efficient and better to be run hundreds of times per frame?
Another edit:
I spent about an hour messing around with it on a sandbox website, and I made this small project:
https://www.khanacademy.org/computer-programming/-/6446241383661568
There I run each one every millisecond and see how quickly each one updates to see which appears more efficient. clip() is on top while CanvasPattern is on the bottom. They both appear to be incredibly fast to me, and I feel that no matter which I chose I will have almost exactly the same results. However, clip() does still appear to be a bit faster as far as I can tell.
See for yourself and let me know what you think!
I am learning JavaScript, coming from eight years of Python experience, and as an exercise I made a Mandelbrot renderer that generates the fractal and uses putImageData to update the canvas from left to right one single-pixel column at a time.
I found that with this approach the visible image in the browser only updates when the full-screen calculation is complete (rather than seeing it appear gradually from left to right as I wanted). I understand that it is the expected behaviour, so I decided to add the "scan-line" animation by using requestAnimationFrame. (something like seen here: Christian Stigen Larsen renderer)
My expectation is that lighter calculations should render faster (as the next frame is available sooner) and higher-iteration calculations should render slower. What I found with my implementation is that the scan-line is consistently slow.
I have isolated this issue from anything to do with my Mandelbrot calculations, as this behaviour also happens for the minimal case below. I am running this on Chrome 83 and see the canvas populate very slowly at a constant rate (approx 30 pixels per second) from left to right.
Is my implementation of rAF incorrect or are my expectations wrong? The renderer I linked to uses setTimeout() to animate, but I read that it is a widely discouraged practice these days.
How should I implement a left-to-right scan update of my canvas at the highest available frame-rate (I am not worried about limiting it at the moment)?
EDIT: For clarity, the code below draws a single thin rectangle at every frame request by rAF, and takes exactly the same amount of time to paint the full canvas as a 100-iteration Mandelbrot render.
This suggest to me that the slowness of it is not due to amount of calculations taking place between frames.
const canvas = document.querySelector('.myCanvas');
const width = window.innerWidth;
const height = window.innerHeight;
const ctx = canvas.getContext('2d');
function anim(timestamp, i) {
if (i < width) {
ctx.fillRect(i, 0, 1, height);
window.requestAnimationFrame(function(timestamp) {
anim(timestamp, i + 1);
})
}
}
ctx.fillStyle = 'rgb(0,0,0)';
window.requestAnimationFrame(function(timestamp) {
anim(timestamp, 0);
});
<!DOCTYPE html>
<html lang="en">
<head>
<canvas class="myCanvas">
<p>Add suitable fallback here.</p>
</canvas>
<script src="slow%20rAF.js"></script>
<style>
body {
margin: 0;
overflow: hidden;
}
</style>
<meta charset="UTF-8">
<title>Mandelbrot</title>
</head>
<body>
</body>
</html>
There is no 'faster' requestAnimationFrame.
Generally the answer to this is to separate calcuations from rendering. Ideally the calculations happen off of the main thread and rendering is handled on the main thread. Even without threading you should be able to find a way to do the calculations outside of the RAF, and just have the RAF render the pixels.
I have a very simple question : is it possible to use globalCompositeOperation for only a restricted number of images?
For example, I draw a lot of stuff into my canvas. And on top of everything, and completed unrelated, I want to do some operations on two images (and I just want the result to be displayed, not both images). How can that be done?
For now, doing such operations affects everything that's already drawn underneath.
So a solution I found is doing the operations in another canvas, which I display on top of my main, first canvas. But this looks bad. First, it hits performances. Then, it doesn't feel intuitive. And last, I loose control over the layers : whatever is in my second canvas will always be on top of the first canvas.
This looks like a pretty simple feature, I hope I'm just bad at googling!
Thanks a lot!
Create a offscreen canvas and use it as a work space. You can create many as long as you have the RAM to store them.
To create an offscreen canvas
function createCanvas(w,h){
var canvas = document.createElement("canvas");
canvas.width = w;
canvas.height = h;
canvas.ctx = ctx.getContext("2d");
return canvas;
}
Easy as that and then you just treat it as an image. To draw to another canvas
var offScreenCan = createCanvas(1024,1024);
offScreenCan.ctx.drawImage(myImage); // put something on the canvas
ctx.drawImage(offScreenCan,0,0); // draw that canvas to another
I attach the context to the canvas rather than wrap the canvas in a containing object. I used to worry that the canvas would affect performance now I convert all images to canvas as soon as I load them and it does not impact performance at all.
I'm trying to accomplish an effect similar to what you might see on the cartoon Chowder (example link) , where shapes serve as masking layers for a texture underneath that stays static. I've begun playing around with this idea by creating a render loop that clears the canvas, saves it's state, then draws a rectangular clipping region, followed by drawing the background texture that occupies the entire width and height of the canvas.
Here's the draw function:
function draw()
{
context.clearRect(0,0, 640, 480);
context.save();
x += velocityX;
y += velocityY;
context.rect(x, y, 40, 40);
context.clip();
context.drawImage(image, 0,0, 640, 480);
context.restore();
}
Basically it just runs at 60 frames per second, updating the position of the rectangle and clipping a background image inside the clipping region. (I know the code isn't structured perfectly, but I was just experimenting to see if this effect was even possible on the canvas).
http://jsfiddle.net/JERje/86/
The problem I seem to be having is that the clipping area from the previous iteration of the loop hangs around creating the weird effect that you see in the fiddle above. I've tried reordering everything in the draw() step of the loop, but the only thing that seems to work is the canvas.width = canvas.width trick for clearing the screen. I'd like to avoid this method of clearing the screen, since it doesn't seem to work in IE, and it also destroys the canvas state. clearRect() should work to clear the screen. What am I doing wrong?
You're using the same HTML5 Canvas paperback I am aren't you.
If you set up an adhoc canvas as I did on your jsfiddle like so:
var newCanvas = document.createElement('canvas');
newCanvas.getContext("2d").drawImage(image,0,0);
A function such as this would be able to hack a section out of that canvas:
context.putImageData(newCanvas.getContext("2d").getImageData(x,y,40,40),x,y);
Thus giving you the chowder effect. Good show man, good luck. Pst me if it doesn't work
EDIT: However this solution will ignore some context scaling transformations. Just be smart about how you handle scale on your own (and you really should be anyways if you want the true "chowder" effect)
So, feel pretty dumb about this, but apparently when you call rect() you also have to make sure to call closePath afterwards in order to close the clipping area. Glad I figured it out finally, now on to adding multiple layers!
Here's the working fiddle: http://jsfiddle.net/JERje/129/