Breaking Down the WebGL Triangle Setup
Breaking Down the WebGL Triangle Setup
WebGL has a track record of being one of Javascript’s more complex API’s. As a web developer intrigued by all that’s interactive, I decided to dive into WebGL’s MDN documentation. I’ve worked with Three.js which abstracts a lot of the hardships of WebGL but I couldn’t help but open up the hood!
The goal of this deep dive was to see if I could make sense of the documentation enough to explain it in simpler terms. Disclaimer — I’ve worked with Three.js and have a bit of knowledge in 3D graphic terminology and patterns. I’ll be sure to explain these concepts if they apply in any way.
To start, we’re going to focus on the creation of a triangle. The reason for this is to establish an understanding of the parts needed to set up webGL.
To get a feel of what will be covered, here are the steps we’ll be going through.
- Set up the HTML canvas
- Get the WebGL context
- Clear the canvas color and set a new one
- Create an array of triangle (x,y )coordinates
- Add vertex and fragment shader code
- Process and compile the shader code
- Create a webGL program
- Create and bind buffers to the webGL program
- Use the program
- Link the GPU information to the CPU
- Draw out the triangle
Setup HTML Canvas
- Create a folder called “RenderTriangle”
- In that folder create an index.html and main.js file
Within the index.html file add the following code:
index.js
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Document</title>
</head>
<body>
<canvas id="canvas"></canvas>
</body>
<script src="main.js"></script>
</html>
- is the entry point for rendering the WebGL context
- We set up the basic HTML code and link the main.js file.
- In the main.js file we will access the canvas id to render webGL content.
Prepare the HTML Canvas
Within the main.js file add the following code which prepares the HTML canvas:
main.js
// get canvas
const canvas = document.getElementById("canvas");
// set width and height of canvas
canvas.width = window.innerWidth;
canvas.height = window.innerHeight;
// get the webgl context
const gl = canvas.getContext("webgl2");
- Get the HTML canvas by id and store it in a variable called “canvas” (you can use any name).
- Set the canvas.width and canvas.height properties of the canvas by accessing the window.innerWidth and window.innerHeight. This sets the rendered display to the size of the browser window.
- Get the WebGL context using canvas.getContext(“webgl2”) and store it in a variable called “gl”.
Clear Color
main.js
gl.clearColor(0.1, 0.2, 0.3, 1.0);
gl.clear(gl.DEPTH_BUFFER_BIT | gl.COLOR_BUFFER_BIT);
Before writing any webGL program you need to set the background color of your canvas. WebGL has two methods for making this happen.
clearColor() — is a method that sets a specific background color. It’s called “clearColor” because when you render WebGL into the HTML canvas, CSS sets the background to the color black. When you call clearColor(), it clears the default color and sets whatever color you want. We can see this below.
Note: The clearColor() has 4 parameters (r, g, b, a)
clear() — after clearColor() is called and an explicit background color is set,
**clear()** must be called to “clear” or reset the buffers to the preset values (the buffers are temporary storage for color and depth information).
The clear() method is one of the drawing methods which means it’s the method that actually renders the color. Without calling the clear() method the canvas won’t show the clearColor. The clear() parameter options are,
are gl.COLOR_BUFFER_BIT
, gl.DEPTH_BUFFER_BIT
, or gl.STENCIL_BUFFER_BIT
.
In the code below you can add multiple parameters to reset in different scenarios.
- gl.DEPTH_BUFFER_BIT — indicates buffers for pixel depth information
- gl.COLOR_BUFFER_BIT — indicates the buffers for pixel color information
Set Triangle Coordinates
main.js
// set position coordinates for shape
const triangleCoords = [0.0, -1.0, 0.0, 1.0, 1.0, 1.0];
Once the the background is set we can set the coordinates needed to create the triangle. The coordinates are stored in an array as (x, y) coordinates.
The below array holds 3-point coordinates. These points connect to form the triangle.
0.0, -1.0
0.0 , 1.0
1.0, 1.0
Adding Vertex and Fragment Shaders
main.js
const vertexShader = `#version 300 es
precision mediump float;
in vec2 position;
void main() {
gl_Position = vec4(position.x, position.y, 0.0, 1.0); //x,y,z,w
}
`;
const fragmentShader = `#version 300 es
precision mediump float;
out vec4 color;
void main () {
color = vec4(0.0,0.0,1.0,1.0); //r,g,b,a
}
`;
After creating a variable for the triangle coordinates, we can set up the shaders.
A shader is a program written in OpenGL ES Shading Language. The program takes position and color information about each vertice point. This information is what’s needed to render geometry.
There are two types of shaders functions that are needed to draw webgl content, the vertex shader and fragment shader
*vertex shader *— The vertex shader function uses position information to render each pixel. Per every render, the vertex shader function runs on each vertex. The vertex shader then transforms each vertex from it’s original coordinates to WebGL coordinates. Each transformed vertex is then saved to the gl_Position variable in the vertex shader program.
*fragment shader *— The fragment shader function is called once for every pixel on a shape to be drawn. This occurs after the vertex shader runs. The fragment shader determines how the color of each pixel and “texel (pixel within a texture)” should be applied. The fragment shader color is saved in the gl_FragColor variable in the fragment shader program.
In the code above we are creating both a vertexShader and fragmentShader constant and storing the shader code in them. The Book of Shaders is a great resource for learning how to write GLSL code.
Processing the Vertex and Fragment Shaders
main.js
// process vertex shader
const shader = gl.createShader(gl.VERTEX_SHADER);
gl.shaderSource(shader, vertexShader);
gl.compileShader(shader);
if (!gl.getShaderParameter(shader, gl.COMPILE_STATUS)) {
console.log(gl.getShaderInfoLog(vertexShader));
}// process fragment shader
const shader = gl.createShader(gl.FRAGMENT_SHADER);
gl.shaderSource(shader, fragmentShader);
gl.compileShader(shader);
if (!gl.getShaderParameter(shader, gl.COMPILE_STATUS)) {
console.log(gl.getShaderInfoLog(fragmentShader));
}
Now that we wrote the shader code (GLSL), we need to create the shader. The shader code still needs to compile. To do this we call the following functions:
createShader() — This creates the shader within the WebGL context
shaderSource() — This takes the GLSL source code that we wrote and sets it into the webGLShader object that was created with createShader.
compileShader() — This compiles the GLSL shader program into data for the WebGLProgram.
The code above processes the vertex and fragment shaders to eventually compile into the WebGLProgram.
Note: An if conditional is added to check if both shaders have compiled properly. If not, an info log will appear. Debugging can be tricky in WebGL so adding these checks is a must.
Creating a WebGL Program
main.js
const program = gl.createProgram();
// Attach pre-existing shaders
gl.attachShader(program, vertexShader);
gl.attachShader(program, fragmentShader);
gl.linkProgram(program);
if (!gl.getProgramParameter(program, gl.LINK_STATUS)) {
const info = gl.getProgramInfoLog(program);
throw "Could not compile WebGL program. \n\n${info}";
}
Let’s review the code above:
After compiling the vertexShader and fragmentShader, we can now create a WebGLProgram. A WebGLProgram is an object that holds the compiled vertexShader and fragmentShader.
createProgram() — Creates and initializes the WebGLProgram
attachShader() — This method attaches the shader to the webGLProgram
linkProgram() — This method links the program object with the shader objects
Lastly, we need to make a conditional check to see if the program is running properly. We do this with the gl.getProgramParameter.
Create and Bind the Buffers
Now that the WebGL Program is created and the shader programs are linked to it, it’s time to create the buffers. What are buffers
?
To simplify it as much as possible, buffers are objects that store vertices and colors. Buffers don’t have any methods or properties that are accessible. Instead, the WebGL context has it’s own methods for handling buffers.
Buffer — “a temporary storage location for data that’s being moved from one place to another”
-wikipedia
We need to create a buffer so that we can store our triangle colors and vertices.
To do this we add the following:
main.js
const buffer = gl.createBuffer();
gl.bindBuffer(gl.ARRAY_BUFFER, buffer);
gl.bufferData(gl.ARRAY_BUFFER, new Float32Array(triangleCoords), gl.STATIC_DRAW);
gl.bindBuffer(gl.ARRAY_BUFFER, null);
In the code above we’re creating a buffer (or temporary storage object) using the createBuffer() method. Then we store it in a constant. Now that we have a buffer object we need to bind it to a target. There are several different target but since we’re storing coordinates in an array we will be using the gl.ARRAY_BUFFER.
To bind the buffer to a target, we use gl.bindBuffer() and pass in the gl.ARRAY_BUFFER (target) and the buffer itself as parameters.
The next step would be to use gl.bufferData() which creates the data store. gl.bufferData() takes the following parameters:
target — gl.ARRAY_BUFFER
data — new Float32Array(triangleCoords)
usage (draw type) — gl.STATIC_DRAW
Lastly, we unbind the buffer from the target to reduce side effects.
Use Program
main.js
gl.useProgram(program);
Once the buffer creation and binding is complete, we can now call the method that sets the WebGLProgram to the rendering state.
Link GPU and CPU
As we get closer to the final step, we need to talk about attributes.
In WebGL, vertex data is stored in a special variable called attributes. attributes are only available to the javascript program and the vertex shader. To access the attributes variable we need to first get the location of the attributes from the GPU. The GPU uses an index to reference the location of the attributes.
main.js
// get index that holds the triangle position information
const position = gl.getAttribLocation(obj.program, obj.gpuVariable);
gl.enableVertexAttribArray(position);
gl.bindBuffer(gl.ARRAY_BUFFER, buffer);
gl.vertexAttribPointer(position, 2, gl.FLOAT, obj.normalize, obj.stride, obj.offset);
Let’s review the code above:
Since we’re rendering a triangle we need the index of the position variable that we set in the vertex shader. We do this using the gl.getAttribLocation() method. By passing in the WebGL program and the position variable name (from the vertex shader) we can get the position attribute’s index.
Next, we need to use the gl.enableVertexAttribArray() method and pass in the position index that we just obtained. This will enable the attributes` storage so we can access it.
We will then rebind our buffer using gl.bindBuffer() and pass in the gl.ARRAY_BUFFER and buffer parameters (the same as when we created the buffers before). Remember in the “Create and Bind the Buffers” section we set the buffer to null to avoid side effects.
When we binded the buffer to the gl.ARRAY_BUFFER we are now able to store our attributes in a specific order. gl.vertexAttribPointer() allows us to do that.
By using gl.vertexAttribPointer() we can pass in the attributes we’d like to store in a specific order. The parameters are ordered first to last.
Sometimes 3D geometry already has a certain format in which the geometry information is set. vertexAttribPointer comes in handy if you need to make modifications to how that geometry information is organized.
Draw Triangle
main.js
gl.drawArrays(gl.TRIANGLES, 0, 3);
Lastly, we can use the gl.drawArrays method to render the triangle. There are other draw methods, but since our vertices are in an array, this method should be used.
gl.drawArrays() takes three parameters:
mode — which specifies the type of primitive to render. (in this case were rendering a triangle)
first — specifies the starting index in the array of vector points (our triangle coordinates). In this case it’s 0.
count — specifies the number of indices to be rendered. ( since it's a triangle we’re rendering 3 indices)
Note: For more complex geometry with a lot of vertices you can use **triangleCoords.length / 2 **to ****get how many indices your geometry has.
Finally, your triangle should be rendered to the screen! Let’s review the steps.
- Set up the HTML canvas
- Get the WebGL context
- Clear the canvas color and set a new one
- Create an array of triangle (x,y )coordinates
- Add vertex and fragment shader code
- Process and compile the shader code
- Create a webGL program
- Create and bind buffers to the webGL program
- Use the program
- Link the GPU information to the CPU
- Draw out the triangle
The API is a complex one so there’s still a lot to learn but understanding this setup has given me a better foundation.
// Set up the HTML canvas
const canvas = document.getElementById("canvas");
canvas.width = window.innerWidth;
canvas.height = window.innerHeight;
// get the webgl context
const gl = canvas.getContext("webgl2");
// Clear the canvas color and set a new one
gl.clearColor(0.1, 0.2, 0.3, 1.0);
gl.clear(gl.DEPTH_BUFFER_BIT | gl.COLOR_BUFFER_BIT);
// Create an array of triangle (x,y )coordinates
const triangleCoords = [0.0, -1.0, 0.0, 1.0, 1.0, 1.0];
// Add vertex and fragment shader code
const vertexShader = `#version 300 es
precision mediump float;
in vec2 position;
void main() {
gl_Position = vec4(position.x, position.y, 0.0, 1.0); //x,y,z,w
}
`;
const fragmentShader = `#version 300 es
precision mediump float;
out vec4 color;
void main () {
color = vec4(0.0,0.0,1.0,1.0); //r,g,b,a
}
`;
// Process and compile the shader code
const vShader = gl.createShader(gl.VERTEX_SHADER);
gl.shaderSource(vShader, vertexShader);
gl.compileShader(vShader);
if (!gl.getShaderParameter(vShader, gl.COMPILE_STATUS)) {
console.log(gl.getShaderInfoLog(vertexShader));
}
const fShader = gl.createShader(gl.FRAGMENT_SHADER);
gl.shaderSource(fShader, fragmentShader);
gl.compileShader(fShader);
if (!gl.getShaderParameter(fShader, gl.COMPILE_STATUS)) {
console.log(gl.getShaderInfoLog(fragmentShader));
}
// Create a webGL program
const program = gl.createProgram();
// Link the GPU information to the CPU
gl.attachShader(program, vShader);
gl.attachShader(program, fShader);
gl.linkProgram(program);
if (!gl.getProgramParameter(program, gl.LINK_STATUS)) {
const info = gl.getProgramInfoLog(program);
throw "Could not compile WebGL program. \n\n${info}";
}
// Create and bind buffers to the webGL program
const buffer = gl.createBuffer();
gl.bindBuffer(gl.ARRAY_BUFFER, buffer);
gl.bufferData(gl.ARRAY_BUFFER, new Float32Array(triangleCoords), gl.STATIC_DRAW);
gl.bindBuffer(gl.ARRAY_BUFFER, null);
// Use the program
gl.useProgram(program);
// Link the GPU information to the CPU
const position = gl.getAttribLocation(program, "position");
gl.enableVertexAttribArray(position);
gl.bindBuffer(gl.ARRAY_BUFFER, buffer);
gl.vertexAttribPointer(position, 2, gl.FLOAT, gl.FALSE, 0, 0);
// render triangle
gl.drawArrays(gl.TRIANGLES, 0, 3);
Here are some invaluable references to help understand this material better.
https://developer.mozilla.org/en-US/docs/Web/API/WebGL_API
https://www.udemy.com/course/webgl-internals/
https://webglfundamentals.org/