12.5 - GLSL Built-in Functions and Variables

Built-in Functions

GLSL provides a significant number of built-in functions and you should be familiar with them. A complete list of these functions can be found on page four of this Quick Reference Card. Please note the following about the functions.

  • The majority of the functions only work on floating point scalars or vectors. They do not work on integers or booleans. The exception is a group of “Vector Relational Functions” that accept Boolean and integer inputs and return Boolean values.

  • The functions are overloaded to accept various inputs and return a result of the same data type. For example, consider the function prototype T sin(T angle), where T represents the data types float, vec2, vec3, or vec4. This represents four versions of the sin function. Each version performs a component-wise sine calculation and returns a result that is the same size as its input.

    float function sin(float angle);
    vec2  function sin(vec2 angles);
    vec3  function sin(vec3 angles);
    vec4  function sin(vec4 angles);

Please review the list of built-in functions on page four of the Quick Reference Card.

Built-in Variables

Shader programs communicate with the graphics pipeline using pre-defined input and output variables.

A Vertex Shader‘s Outputs

A vertex shader has two outputs:

  • gl_Position: a vec4 position of a vertex in “clip coordinates”. Clip coordinates were described in detail in section 8.
    • The values for x and y are in the range [-1.0,+1.0] and represent the location of a vertex in the viewing window.
    • The z value is in the range [-1.0,+1.0] and represents the distance of the vertex from the camera.
    • The w value is 1.0 for orthogonal projections, or the w value is the perspective divide value for perspective projections.
    • Any vertex outside the clipping cube is clipped to the cube’s boundaries.
  • gl_PointSize: the number of pixels to use to render a point. It is a float value that can have a fractional part. It only applies to the rendering of single points, not to the vertices of lines and triangles. If no value is specified, its default value it 1.0.

The outputs of a vertex shader are used by the graphics pipeline to determine the pixels (i.e., the fragments) in the viewing window that compose a graphics primitive. The rasterizer in the graphics pipeline creates a fragment for each pixel, calculates the interpolated values for any varying variables, and sets the values for the following fragment shader input values.

Inputs to a Fragment Shader

  • gl_FragCoord: a vec4 value that holds the (x,y,z,w) value of the fragment. This is the value of gl_Position after it has been transformed by the viewport transform and the perspective divide has been performed. Therefore, the (x,y) values are the location of the fragment in the image to be rendered. The z value is the distance from the camera. Note that these are floating point values and that the (x,y) values are the center of a pixel. For example, the bottom-left corner pixel has an (x,y) value of (0.5, 0.5).
  • gl_FrontFacing: a Boolean value that is true if this fragment is part of a front-facing primitive. This only applies to triangles. A triangle is “front facing” if its normal vector is pointing toward the camera. The normal vector is calculated from the triangle’s vertices, and a counter-clockwise ordering of the vertices will produce a “front-facing” normal vector.
  • gl_PointCoord: is a vec2 value where its (x,y) values are in the range [0.0,1.0]. The values indicate the relative location of this fragment within the rendering of a point. The point’s total size, in pixels, comes from the gl_PointSize output variable. The origin of the coordinate system for a point rendering is the upper left corner of the square that covers the point. (This only applies to rendered points, gl.POINTS. It is undefined when rendering gl.LINES and gl.TRIANGLES.)

Outputs from a Fragment Shader

A WebGL fragment shader has one output – a color value for its fragment.

  • gl_FragColor : a RGBA (vec4) value that is placed into the color buffer for the fragment it is processing.

OpenGL supports rendering to multiple color buffers for the creation of advanced visual effects. Writing colors to gl_FragData[n] allows a fragment shader to create multiple images that can be “composited” into a single final rendering. WebGL 1.0 does not support multiple color buffers without using extensions, so we will not discuss them here.

Next Section - 12.6 - GLSL Compiling and Linking