Built-in Shader Uniforms

PlayCanvas supplies the following numerical uniforms when rendering meshes in forward passes. Use them in ShaderMaterial shaders and custom material chunks without calling Material.setParameter() for these names. To set the values of uniforms yourself, see Shader Parameters.

Camera values follow the camera or XR view being rendered. The list describes forward rendering; do not assume the same inputs are populated in shadow, picking, compute, or standalone post-processing passes.

Available Uniforms

The uniforms can be read in vertex and fragment shaders. Declare the names you use, with the types shown below. In WGSL, access them through the uniform. prefix.

Uniform	GLSL type	WGSL type	Value
matrix_view	mat4	mat4x4f	Transforms world-space positions to view space.
matrix_viewInverse	mat4	mat4x4f	Inverse of matrix_view; transforms view-space positions to world space.
matrix_view3	mat3	mat3x3f	Upper-left 3×3 part of matrix_view. Transforms directions without translation.
matrix_projection	mat4	mat4x4f	Transforms view-space positions to clip space. Includes the renderer's projection adjustments, such as jitter, render-target Y flip, and the graphics backend's depth convention.
matrix_viewProjection	mat4	mat4x4f	matrix_projection * matrix_view; transforms world-space positions to clip space.
view_position	vec3	vec3f	Camera position in world space; the current eye's position in XR.
camera_params	vec4	vec4f	Camera clip planes and projection type: 1 / farClip, farClip, nearClip, and 1 for an orthographic camera or 0 for a perspective one. Use it to linearize depth values.
screen_size	vec4	vec4f	Canvas drawing-buffer width, height, inverse width, and inverse height. These are pixel dimensions, not CSS dimensions.
viewport_size	vec4	vec4f	Camera viewport width, height, inverse width, and inverse height in pixels. Accounts for the render target, camera.rect, and the XR eye viewport. Contains no viewport origin.
view_index	uint	u32	Zero-based index of the XR view currently being rendered. Use it to select per-view data such as a depth texture array layer.

The projection matrices are the versions used for rendering. They can differ from a projection matrix calculated in application code before jitter or backend adjustments. Prefer matrix_viewProjection when projecting a world-space position in a shader.

matrix_model and matrix_normal are per-mesh inputs, separate from these camera and scene values. For transformations that also support skinning, morphing, and instancing, use the vertex shader chunks.

The scene depth and color textures, uSceneDepthMap and uSceneColorMap, are described in Depth Layer.

Declaring and Reading Uniforms

For example, convert a world-space direction to view space:

GLSL

uniform mat3 matrix_view3;

vec3 directionToView(vec3 worldDirection) {
    return matrix_view3 * worldDirection;
}

WGSL

uniform matrix_view3: mat3x3f;

fn directionToView(worldDirection: vec3f) -> vec3f {
    return uniform.matrix_view3 * worldDirection;
}

For the full declaration rules, see GLSL Specifics and WGSL Reflection.

Screen Size and Viewport Size

Both size uniforms use the layout [width, height, 1 / width, 1 / height], but describe different areas:

• screen_size always describes the canvas drawing buffer. Rendering to a smaller offscreen target does not change it.
• viewport_size describes the camera's viewport within its current target. It changes for offscreen targets, partial camera rectangles, and XR eye viewports.

For a 1920×1080 canvas with a camera rendering into a 960×540 target, screen_size.xy is (1920, 1080) and viewport_size.xy is (960, 540) when the camera fills the target. If the camera occupies half the target's width, viewport_size.xy is (480, 540).

Do not use screen_size as a general texture-size uniform. To normalize framebuffer coordinates for a partial viewport, also account for the viewport origin; neither size uniform supplies that origin.