Struct Mat4 

#[repr(C)]
pub struct Mat4 {
    pub x_axis: Vec4,
    pub y_axis: Vec4,
    pub z_axis: Vec4,
    pub w_axis: Vec4,
}

A 4x4 column major matrix.

If you are primarily dealing with 3D affine transformations considering using Affine3A which is faster than a 4x4 matrix for some affine operations.

Affine transformations including 3D translation, rotation and scale can be created using methods such as Self::from_translation(), Self::from_quat(), Self::from_scale() and Self::from_scale_rotation_translation().

The Self::transform_point3() and Self::transform_vector3() convenience methods are provided for performing affine transformations on 3D vectors and points. These multiply 3D inputs as 4D vectors with an implicit w value of 1 for points and 0 for vectors respectively. These methods assume that Self contains a valid affine transform.

SIMD vector types are used for storage on supported platforms.

This type is 16 byte aligned.

Fields

x_axis: Vec4

y_axis: Vec4

z_axis: Vec4

w_axis: Vec4

Implementations

impl Mat4

pub const ZERO: Mat4

A 4x4 matrix with all elements set to 0.0.

pub const IDENTITY: Mat4

A 4x4 identity matrix, where all diagonal elements are 1, and all off-diagonal elements are 0.

pub const NAN: Mat4

All NAN:s.

pub const fn from_cols( x_axis: Vec4, y_axis: Vec4, z_axis: Vec4, w_axis: Vec4, ) -> Mat4

Creates a 4x4 matrix from four column vectors.

See also Self::from_rows when the data is in row major order.

pub const fn from_rows(row0: Vec4, row1: Vec4, row2: Vec4, row3: Vec4) -> Mat4

Creates a 4x4 matrix from four row vectors.

Matrices are stored in column major order, so the given rows are permuted into the matrix layout. Use Self::from_cols instead when the data is already in column major order.

pub const fn from_cols_array(m: &[f32; 16]) -> Mat4

Creates a 4x4 matrix from a [f32; 16] array stored in column major order.

If the data is in row major order use Self::from_rows_array instead.

pub const fn to_cols_array(&self) -> [f32; 16]

Creates a [f32; 16] array storing data in column major order.

If you require the data in row major order use Self::to_rows_array instead.

pub const fn from_cols_array_2d(m: &[[f32; 4]; 4]) -> Mat4

Creates a 4x4 matrix from a [[f32; 4]; 4] 4D array stored in column major order.

If the data is in row major order transpose the returned matrix.

Examples found in repository

examples/3d/irradiance_volumes.rs (lines 59-64)

static VOXEL_FROM_WORLD: Mat4 = Mat4::from_cols_array_2d(&[
    [-42.317566, 0.0, 0.0, 0.0],
    [0.0, 0.0, 44.601563, 0.0],
    [0.0, 16.73776, 0.0, 0.0],
    [0.0, 6.544792, 0.0, 1.0],
]);

pub const fn to_cols_array_2d(&self) -> [[f32; 4]; 4]

Creates a [[f32; 4]; 4] 4D array storing data in column major order.

If you require row major order transpose the matrix first.

pub const fn from_rows_array(m: &[f32; 16]) -> Mat4

Creates a 4x4 matrix from a [f32; 16] array stored in row major order.

Matrices are stored in column major order, so the array is permuted into the matrix layout. Use Self::from_cols_array instead when the data is already in column major order.

pub const fn to_rows_array(&self) -> [f32; 16]

Creates a [f32; 16] array storing data in row major order.

Matrices are stored in column major order, so the array is permuted out of the column major storage. Use Self::to_cols_array instead when you want data in column major order.

pub const fn from_diagonal(diagonal: Vec4) -> Mat4

Creates a 4x4 matrix with its diagonal set to diagonal and all other entries set to 0.

pub fn from_scale_rotation_translation( scale: Vec3, rotation: Quat, translation: Vec3, ) -> Mat4

Creates an affine transformation matrix from the given 3D scale, rotation and translation.

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

Panics

Will panic if rotation is not normalized when glam_assert is enabled.

pub fn from_rotation_translation(rotation: Quat, translation: Vec3) -> Mat4

Creates an affine transformation matrix from the given 3D translation.

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

Panics

Will panic if rotation is not normalized when glam_assert is enabled.

pub fn to_scale_rotation_translation(&self) -> (Vec3, Quat, Vec3)

Extracts scale, rotation and translation from self. The input matrix is expected to be a 3D affine transformation matrix otherwise the output will be invalid.

Panics

Will panic if self is not a valid affine transformation matrix, if the determinant of the 3x3 linear part (the rotation and scale part of the transform) is zero, when glam_assert is enabled.

pub fn from_quat(rotation: Quat) -> Mat4

Creates an affine transformation matrix from the given rotation quaternion.

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

Panics

Will panic if rotation is not normalized when glam_assert is enabled.

pub fn from_mat3(m: Mat3) -> Mat4

Creates an affine transformation matrix from the given 3x3 linear transformation matrix.

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

pub fn from_mat3_translation(mat3: Mat3, translation: Vec3) -> Mat4

Creates an affine transformation matrics from a 3x3 matrix (expressing scale, shear and rotation) and a translation vector.

Equivalent to Mat4::from_translation(translation) * Mat4::from_mat3(mat3)

pub fn from_mat3a(m: Mat3A) -> Mat4

Creates an affine transformation matrix from the given 3x3 linear transformation matrix.

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

Examples found in repository

examples/3d/mirror.rs (line 355)

fn calculate_mirror_camera_transform_and_projection(
    main_camera_transform: &Transform,
    main_camera_projection: &PerspectiveProjection,
    mirror_transform: &Transform,
) -> (Transform, PerspectiveProjection) {
    // Calculate the reflection matrix (a.k.a. Householder matrix) that will
    // reflect the scene across the mirror plane.
    //
    // Note that you must calculate this in *matrix* form and only *afterward*
    // convert to a `Transform` instead of composing `Transform`s. This is
    // because the reflection matrix has non-uniform scale, and composing
    // transforms can't always handle composition of matrices with non-uniform
    // scales.
    let mirror_camera_transform = Transform::from_matrix(
        Mat4::from_mat3a(reflection_matrix(Vec3::NEG_Z)) * main_camera_transform.to_matrix(),
    );

    // Compute the distance from the camera to the mirror plane. This will be
    // used to calculate the distance to the near clip plane for the mirror
    // world.
    let distance_from_camera_to_mirror = InfinitePlane3d::new(mirror_transform.rotation * Vec3::Y)
        .signed_distance(
            Isometry3d::IDENTITY,
            mirror_transform.translation - main_camera_transform.translation,
        );

    // Compute the normal of the mirror plane in view space.
    let view_from_world = main_camera_transform.compute_affine().matrix3.inverse();
    let mirror_projection_plane_normal =
        (view_from_world * (mirror_transform.rotation * Vec3::NEG_Y)).normalize();

    // Compute the final projection. It should match the main camera projection,
    // except that `near` and `near_normal` should be set to the updated near
    // plane and near normal plane as above.
    let mirror_camera_projection = PerspectiveProjection {
        near_clip_plane: mirror_projection_plane_normal.extend(distance_from_camera_to_mirror),
        ..*main_camera_projection
    };

    (mirror_camera_transform, mirror_camera_projection)
}

pub fn from_translation(translation: Vec3) -> Mat4

Creates an affine transformation matrix from the given 3D translation.

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

Examples found in repository

tests/3d/test_invalid_skinned_mesh.rs (line 145)

fn setup_meshes(
    mut commands: Commands,
    mut mesh_assets: ResMut<Assets<Mesh>>,
    mut material_assets: ResMut<Assets<StandardMaterial>>,
    mut inverse_bindposes_assets: ResMut<Assets<SkinnedMeshInverseBindposes>>,
) {
    // Create a mesh with two rectangles.
    let unskinned_mesh = Mesh::new(
        PrimitiveTopology::TriangleList,
        RenderAssetUsages::default(),
    )
    .with_inserted_attribute(
        Mesh::ATTRIBUTE_POSITION,
        vec![
            [-0.3, -0.3, 0.0],
            [0.3, -0.3, 0.0],
            [-0.3, 0.3, 0.0],
            [0.3, 0.3, 0.0],
            [-0.4, 0.8, 0.0],
            [0.4, 0.8, 0.0],
            [-0.4, 1.8, 0.0],
            [0.4, 1.8, 0.0],
        ],
    )
    .with_inserted_attribute(Mesh::ATTRIBUTE_NORMAL, vec![[0.0, 0.0, 1.0]; 8])
    .with_inserted_indices(Indices::U16(vec![0, 1, 3, 0, 3, 2, 4, 5, 7, 4, 7, 6]));

    // Copy the mesh and add skinning attributes that bind each rectangle to a joint.
    let skinned_mesh = unskinned_mesh
        .clone()
        .with_inserted_attribute(
            Mesh::ATTRIBUTE_JOINT_INDEX,
            VertexAttributeValues::Uint16x4(vec![
                [0, 0, 0, 0],
                [0, 0, 0, 0],
                [0, 0, 0, 0],
                [0, 0, 0, 0],
                [1, 0, 0, 0],
                [1, 0, 0, 0],
                [1, 0, 0, 0],
                [1, 0, 0, 0],
            ]),
        )
        .with_inserted_attribute(
            Mesh::ATTRIBUTE_JOINT_WEIGHT,
            vec![[1.00, 0.00, 0.0, 0.0]; 8],
        );

    let unskinned_mesh_handle = mesh_assets.add(unskinned_mesh);
    let skinned_mesh_handle = mesh_assets.add(skinned_mesh);

    let inverse_bindposes_handle = inverse_bindposes_assets.add(vec![
        Mat4::IDENTITY,
        Mat4::from_translation(Vec3::new(0.0, -1.3, 0.0)),
    ]);

    let mesh_material_handle = material_assets.add(StandardMaterial::default());

    let background_material_handle = material_assets.add(StandardMaterial {
        base_color: Color::srgb(0.05, 0.15, 0.05),
        reflectance: 0.2,
        ..default()
    });

    #[derive(PartialEq)]
    enum Variation {
        Normal,
        MissingMeshAttributes,
        MissingJointEntity,
        MissingSkinnedMeshComponent,
    }

    for (index, variation) in [
        Variation::Normal,
        Variation::MissingMeshAttributes,
        Variation::MissingJointEntity,
        Variation::MissingSkinnedMeshComponent,
    ]
    .into_iter()
    .enumerate()
    {
        // Skip variations that are currently broken. See https://github.com/bevyengine/bevy/issues/16929,
        // https://github.com/bevyengine/bevy/pull/18074.
        if (variation == Variation::MissingSkinnedMeshComponent)
            || (variation == Variation::MissingMeshAttributes)
        {
            continue;
        }

        let transform = Transform::from_xyz(((index as f32) - 1.5) * 4.5, 0.0, 0.0);

        let joint_0 = commands.spawn(transform).id();

        let joint_1 = commands
            .spawn((ChildOf(joint_0), AnimatedJoint, Transform::IDENTITY))
            .id();

        if variation == Variation::MissingJointEntity {
            commands.entity(joint_1).despawn();
        }

        let mesh_handle = match variation {
            Variation::MissingMeshAttributes => &unskinned_mesh_handle,
            _ => &skinned_mesh_handle,
        };

        let mut entity_commands = commands.spawn((
            Mesh3d(mesh_handle.clone()),
            MeshMaterial3d(mesh_material_handle.clone()),
            transform,
        ));

        if variation != Variation::MissingSkinnedMeshComponent {
            entity_commands.insert(SkinnedMesh {
                inverse_bindposes: inverse_bindposes_handle.clone(),
                joints: vec![joint_0, joint_1],
            });
        }

        // Add a square behind the mesh to distinguish it from the other meshes.
        commands.spawn((
            Transform::from_xyz(transform.translation.x, transform.translation.y, -0.8),
            Mesh3d(mesh_assets.add(Plane3d::default().mesh().size(4.3, 4.3).normal(Dir3::Z))),
            MeshMaterial3d(background_material_handle.clone()),
        ));
    }
}

tests/3d/test_skinned_mesh_bounds.rs (line 188)

fn spawn_custom_meshes(
    mut commands: Commands,
    mut mesh_assets: ResMut<Assets<Mesh>>,
    mut material_assets: ResMut<Assets<StandardMaterial>>,
    mut inverse_bindposes_assets: ResMut<Assets<SkinnedMeshInverseBindposes>>,
) {
    let mesh_handle = mesh_assets.add(
        Mesh::new(
            PrimitiveTopology::TriangleStrip,
            // Test that skinned mesh bounds work even if the mesh is render
            // world only.
            RenderAssetUsages::RENDER_WORLD,
        )
        .with_inserted_attribute(
            Mesh::ATTRIBUTE_POSITION,
            vec![
                [-0.5, 0.0, 0.0],
                [0.5, 0.0, 0.0],
                [-0.5, 0.5, 0.0],
                [0.5, 0.5, 0.0],
                [-0.5, 1.0, 0.0],
                [0.5, 1.0, 0.0],
                [-0.5, 1.5, 0.0],
                [0.5, 1.5, 0.0],
                [-0.5, 2.0, 0.0],
                [0.5, 2.0, 0.0],
            ],
        )
        .with_inserted_attribute(Mesh::ATTRIBUTE_NORMAL, vec![[0.0, 0.0, 1.0]; 10])
        .with_inserted_attribute(
            Mesh::ATTRIBUTE_JOINT_INDEX,
            VertexAttributeValues::Uint16x4(vec![
                [1, 0, 0, 0],
                [1, 0, 0, 0],
                [1, 2, 0, 0],
                [1, 2, 0, 0],
                [1, 2, 0, 0],
                [1, 2, 0, 0],
                [2, 1, 0, 0],
                [2, 1, 0, 0],
                [2, 0, 0, 0],
                [2, 0, 0, 0],
            ]),
        )
        .with_inserted_attribute(
            Mesh::ATTRIBUTE_JOINT_WEIGHT,
            vec![
                [1.00, 0.00, 0.0, 0.0],
                [1.00, 0.00, 0.0, 0.0],
                [0.75, 0.25, 0.0, 0.0],
                [0.75, 0.25, 0.0, 0.0],
                [0.50, 0.50, 0.0, 0.0],
                [0.50, 0.50, 0.0, 0.0],
                [0.75, 0.25, 0.0, 0.0],
                [0.75, 0.25, 0.0, 0.0],
                [1.00, 0.00, 0.0, 0.0],
                [1.00, 0.00, 0.0, 0.0],
            ],
        )
        .with_generated_skinned_mesh_bounds()
        .unwrap(),
    );

    let inverse_bindposes_handle = inverse_bindposes_assets.add(vec![
        Mat4::from_translation(Vec3::new(0.0, 0.0, 0.0)),
        Mat4::from_translation(Vec3::new(0.0, 0.0, 0.0)),
        Mat4::from_translation(Vec3::new(0.0, -1.0, 0.0)),
    ]);

    struct MeshInstance {
        animations: [CustomAnimationId; 2],
    }

    let mesh_instances = [
        // Simple cases. First joint is still, second joint is all rotation/translation/scale variations.
        MeshInstance { animations: [0, 1] },
        MeshInstance { animations: [0, 2] },
        MeshInstance { animations: [0, 3] },
        MeshInstance { animations: [0, 4] },
        MeshInstance { animations: [0, 5] },
        MeshInstance { animations: [0, 6] },
        MeshInstance { animations: [0, 7] },
        MeshInstance { animations: [0, 8] },
        // Skewed cases. First joint is non-uniform scaling, second joint is rotation/translation variations.
        MeshInstance { animations: [9, 1] },
        MeshInstance { animations: [9, 2] },
        MeshInstance { animations: [9, 3] },
        MeshInstance { animations: [9, 4] },
        MeshInstance { animations: [9, 5] },
    ];

    for (i, mesh_instance) in mesh_instances.iter().enumerate() {
        let x = ((i as f32) * 2.0) - ((mesh_instances.len() - 1) as f32);

        let base_entity = commands
            .spawn((Transform::from_xyz(x, 0.0, 0.0), Visibility::default()))
            .id();

        let joints = vec![
            commands.spawn((Transform::IDENTITY,)).id(),
            commands
                .spawn((
                    CustomAnimation(mesh_instance.animations[0]),
                    Transform::IDENTITY,
                ))
                .id(),
            commands
                .spawn((
                    CustomAnimation(mesh_instance.animations[1]),
                    Transform::IDENTITY,
                ))
                .id(),
        ];

        commands.entity(joints[0]).insert(ChildOf(base_entity));

        commands.entity(joints[1]).insert(ChildOf(joints[0]));
        commands.entity(joints[2]).insert(ChildOf(joints[1]));

        let mesh_entity = commands
            .spawn((
                Transform::IDENTITY,
                Mesh3d(mesh_handle.clone()),
                MeshMaterial3d(material_assets.add(StandardMaterial {
                    base_color: Color::WHITE,
                    cull_mode: None,
                    ..default()
                })),
                SkinnedMesh {
                    inverse_bindposes: inverse_bindposes_handle.clone(),
                    joints: joints.clone(),
                },
                DynamicSkinnedMeshBounds,
            ))
            .id();

        commands.entity(mesh_entity).insert(ChildOf(base_entity));
    }
}

examples/animation/custom_skinned_mesh.rs (line 53)

fn setup(
    mut commands: Commands,
    asset_server: Res<AssetServer>,
    mut meshes: ResMut<Assets<Mesh>>,
    mut materials: ResMut<Assets<StandardMaterial>>,
    mut skinned_mesh_inverse_bindposes_assets: ResMut<Assets<SkinnedMeshInverseBindposes>>,
) {
    // Create a camera
    commands.spawn((
        Camera3d::default(),
        Transform::from_xyz(2.5, 2.5, 9.0).looking_at(Vec3::ZERO, Vec3::Y),
    ));

    // Create inverse bindpose matrices for a skeleton consists of 2 joints
    let inverse_bindposes = skinned_mesh_inverse_bindposes_assets.add(vec![
        Mat4::from_translation(Vec3::new(-0.5, -1.0, 0.0)),
        Mat4::from_translation(Vec3::new(-0.5, -1.0, 0.0)),
    ]);

    // Create a mesh
    let mesh = Mesh::new(
        PrimitiveTopology::TriangleList,
        RenderAssetUsages::RENDER_WORLD,
    )
    // Set mesh vertex positions
    .with_inserted_attribute(
        Mesh::ATTRIBUTE_POSITION,
        vec![
            [0.0, 0.0, 0.0],
            [1.0, 0.0, 0.0],
            [0.0, 0.5, 0.0],
            [1.0, 0.5, 0.0],
            [0.0, 1.0, 0.0],
            [1.0, 1.0, 0.0],
            [0.0, 1.5, 0.0],
            [1.0, 1.5, 0.0],
            [0.0, 2.0, 0.0],
            [1.0, 2.0, 0.0],
        ],
    )
    // Add UV coordinates that map the left half of the texture since its a 1 x
    // 2 rectangle.
    .with_inserted_attribute(
        Mesh::ATTRIBUTE_UV_0,
        vec![
            [0.0, 0.00],
            [0.5, 0.00],
            [0.0, 0.25],
            [0.5, 0.25],
            [0.0, 0.50],
            [0.5, 0.50],
            [0.0, 0.75],
            [0.5, 0.75],
            [0.0, 1.00],
            [0.5, 1.00],
        ],
    )
    // Set mesh vertex normals
    .with_inserted_attribute(Mesh::ATTRIBUTE_NORMAL, vec![[0.0, 0.0, 1.0]; 10])
    // Set mesh vertex joint indices for mesh skinning.
    // Each vertex gets 4 indices used to address the `JointTransforms` array in the vertex shader
    //  as well as `SkinnedMeshJoint` array in the `SkinnedMesh` component.
    // This means that a maximum of 4 joints can affect a single vertex.
    .with_inserted_attribute(
        Mesh::ATTRIBUTE_JOINT_INDEX,
        // Need to be explicit here as [u16; 4] could be either Uint16x4 or Unorm16x4.
        VertexAttributeValues::Uint16x4(vec![
            [0, 0, 0, 0],
            [0, 0, 0, 0],
            [0, 1, 0, 0],
            [0, 1, 0, 0],
            [0, 1, 0, 0],
            [0, 1, 0, 0],
            [0, 1, 0, 0],
            [0, 1, 0, 0],
            [0, 1, 0, 0],
            [0, 1, 0, 0],
        ]),
    )
    // Set mesh vertex joint weights for mesh skinning.
    // Each vertex gets 4 joint weights corresponding to the 4 joint indices assigned to it.
    // The sum of these weights should equal to 1.
    .with_inserted_attribute(
        Mesh::ATTRIBUTE_JOINT_WEIGHT,
        vec![
            [1.00, 0.00, 0.0, 0.0],
            [1.00, 0.00, 0.0, 0.0],
            [0.75, 0.25, 0.0, 0.0],
            [0.75, 0.25, 0.0, 0.0],
            [0.50, 0.50, 0.0, 0.0],
            [0.50, 0.50, 0.0, 0.0],
            [0.25, 0.75, 0.0, 0.0],
            [0.25, 0.75, 0.0, 0.0],
            [0.00, 1.00, 0.0, 0.0],
            [0.00, 1.00, 0.0, 0.0],
        ],
    )
    // Tell bevy to construct triangles from a list of vertex indices,
    // where each 3 vertex indices form a triangle.
    .with_inserted_indices(Indices::U16(vec![
        0, 1, 3, 0, 3, 2, 2, 3, 5, 2, 5, 4, 4, 5, 7, 4, 7, 6, 6, 7, 9, 6, 9, 8,
    ]))
    // Create skinned mesh bounds. Together with the `DynamicSkinnedMeshBounds`
    // component, this will ensure the mesh is correctly frustum culled.
    .with_generated_skinned_mesh_bounds()
    .unwrap();

    let mesh = meshes.add(mesh);

    // We're seeding the PRNG here to make this example deterministic for testing purposes.
    // This isn't strictly required in practical use unless you need your app to be deterministic.
    let mut rng = ChaCha8Rng::seed_from_u64(42);

    for i in -5..5 {
        // Create joint entities
        let joint_0 = commands
            .spawn(Transform::from_xyz(
                i as f32 * 1.5,
                0.0,
                // Move quads back a small amount to avoid Z-fighting and not
                // obscure the transform gizmos.
                -(i as f32 * 0.01).abs(),
            ))
            .id();
        let joint_1 = commands.spawn((AnimatedJoint(i), Transform::IDENTITY)).id();

        // Set joint_1 as a child of joint_0.
        commands.entity(joint_0).add_children(&[joint_1]);

        // Each joint in this vector corresponds to each inverse bindpose matrix in `SkinnedMeshInverseBindposes`.
        let joint_entities = vec![joint_0, joint_1];

        // Create skinned mesh renderer. Note that its transform doesn't affect the position of the mesh.
        commands.spawn((
            Mesh3d(mesh.clone()),
            MeshMaterial3d(materials.add(StandardMaterial {
                base_color: Color::srgb(
                    rng.random_range(0.0..1.0),
                    rng.random_range(0.0..1.0),
                    rng.random_range(0.0..1.0),
                ),
                base_color_texture: Some(asset_server.load("textures/uv_checker_bw.png")),
                ..default()
            })),
            SkinnedMesh {
                inverse_bindposes: inverse_bindposes.clone(),
                joints: joint_entities,
            },
            DynamicSkinnedMeshBounds,
        ));
    }
}

pub fn from_axis_angle(axis: Vec3, angle: f32) -> Mat4

Creates an affine transformation matrix containing a 3D rotation around a normalized rotation axis of angle (in radians).

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

Panics

Will panic if axis is not normalized when glam_assert is enabled.

pub fn from_euler(order: EulerRot, a: f32, b: f32, c: f32) -> Mat4

Creates a affine transformation matrix containing a rotation from the given euler rotation sequence and angles (in radians).

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

pub fn to_euler(&self, order: EulerRot) -> (f32, f32, f32)

Extract Euler angles with the given Euler rotation order.

Note if the upper 3x3 matrix contain scales, shears, or other non-rotation transformations then the resulting Euler angles will be ill-defined.

Panics

Will panic if any column of the upper 3x3 rotation matrix is not normalized when glam_assert is enabled.

pub fn from_rotation_x(angle: f32) -> Mat4

Creates an affine transformation matrix containing a 3D rotation around the x axis of angle (in radians).

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

pub fn from_rotation_y(angle: f32) -> Mat4

Creates an affine transformation matrix containing a 3D rotation around the y axis of angle (in radians).

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

pub fn from_rotation_z(angle: f32) -> Mat4

Creates an affine transformation matrix containing a 3D rotation around the z axis of angle (in radians).

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

pub fn from_scale(scale: Vec3) -> Mat4

Creates an affine transformation matrix containing the given 3D non-uniform scale.

The resulting matrix can be used to transform 3D points and vectors. See Self::transform_point3() and Self::transform_vector3().

Panics

Will panic if all elements of scale are zero when glam_assert is enabled.

pub const fn from_cols_slice(slice: &[f32]) -> Mat4

Creates a 4x4 matrix from the first 16 values in slice.

See also Self::from_rows_slice when the slice is in row major order.

Panics

Panics if slice is less than 16 elements long.

pub fn write_cols_to_slice(&self, slice: &mut [f32])

Writes the columns of self to the first 16 elements in slice.

Panics

Panics if slice is less than 16 elements long.

pub const fn from_rows_slice(slice: &[f32]) -> Mat4

Creates a 4x4 matrix from the first 16 values in slice, stored in row major order.

Matrices are stored in column major order, so the slice is permuted into the matrix layout. Use Self::from_cols_slice instead when the slice is already in column major order.

Panics

Panics if slice is less than 16 elements long.

pub fn col(&self, index: usize) -> Vec4

Returns the matrix column for the given index.

Panics

Panics if index is greater than 3.

pub fn col_mut(&mut self, index: usize) -> &mut Vec4

Returns a mutable reference to the matrix column for the given index.

Panics

Panics if index is greater than 3.

Examples found in repository

examples/camera/custom_projection.rs (line 25)

    fn get_clip_from_view(&self) -> Mat4 {
        let mut mat = self.perspective.get_clip_from_view();
        mat.col_mut(2)[0] = self.horizontal_obliqueness;
        mat.col_mut(2)[1] = self.vertical_obliqueness;
        mat
    }

    fn get_clip_from_view_for_sub(&self, sub_view: &bevy::camera::SubCameraView) -> Mat4 {
        let mut mat = self.perspective.get_clip_from_view_for_sub(sub_view);
        mat.col_mut(2)[0] = self.horizontal_obliqueness;
        mat.col_mut(2)[1] = self.vertical_obliqueness;
        mat
    }

pub fn row(&self, index: usize) -> Vec4

Returns the matrix row for the given index.

See also Self::set_row when you need to change the row.

Panics

Panics if index is greater than 3.

pub fn set_row(&mut self, index: usize, row: Vec4)

Sets the matrix row for the given index.

Matrices are stored in column major order, so the row is spread across all 4 columns and writing it touches every column. Use Self::col_mut instead when you can work with columns. See also Self::row.

Panics

Panics if index is greater than 3.

pub fn is_finite(&self) -> bool

Returns true if, and only if, all elements are finite. If any element is either NaN, positive or negative infinity, this will return false.

pub fn is_nan(&self) -> bool

Returns true if any elements are NaN.

pub fn transpose(&self) -> Mat4

Returns the transpose of self.

pub fn diagonal(&self) -> Vec4

Returns the diagonal of self.

pub fn determinant(&self) -> f32

Returns the determinant of self.

pub fn inverse(&self) -> Mat4

Returns the inverse of self.

If the matrix is not invertible the returned matrix will be invalid. The returned matrix will also be invalid if the inverse is not finite, which can happen when self contains very large or very small values. Use Self::try_inverse or Self::inverse_or_zero to detect these cases.

Panics

Will panic if the resulting inverted matrix is not finite when glam_assert is enabled.

Examples found in repository

examples/3d/irradiance_volumes.rs (line 557)

fn create_cubes(
    image_assets: Res<Assets<Image>>,
    mut commands: Commands,
    irradiance_volumes: Query<(&IrradianceVolume, &GlobalTransform)>,
    voxel_cube_parents: Query<Entity, With<VoxelCubeParent>>,
    voxel_cubes: Query<Entity, With<VoxelCube>>,
    example_assets: Res<ExampleAssets>,
    mut voxel_visualization_material_assets: ResMut<Assets<VoxelVisualizationMaterial>>,
) {
    // If voxel cubes have already been spawned, don't do anything.
    if !voxel_cubes.is_empty() {
        return;
    }

    let Some(voxel_cube_parent) = voxel_cube_parents.iter().next() else {
        return;
    };

    for (irradiance_volume, global_transform) in irradiance_volumes.iter() {
        let Some(image) = image_assets.get(&irradiance_volume.voxels) else {
            continue;
        };

        let resolution = image.texture_descriptor.size;

        let voxel_cube_material = voxel_visualization_material_assets.add(ExtendedMaterial {
            base: StandardMaterial::from(Color::from(RED)),
            extension: VoxelVisualizationExtension {
                irradiance_volume_info: VoxelVisualizationIrradianceVolumeInfo {
                    world_from_voxel: VOXEL_FROM_WORLD.inverse(),
                    voxel_from_world: VOXEL_FROM_WORLD,
                    resolution: uvec3(
                        resolution.width,
                        resolution.height,
                        resolution.depth_or_array_layers,
                    ),
                    intensity: IRRADIANCE_VOLUME_INTENSITY,
                },
            },
        });

        let scale = vec3(
            1.0 / resolution.width as f32,
            1.0 / resolution.height as f32,
            1.0 / resolution.depth_or_array_layers as f32,
        );

        // Spawn a cube for each voxel.
        for z in 0..resolution.depth_or_array_layers {
            for y in 0..resolution.height {
                for x in 0..resolution.width {
                    let uvw = (uvec3(x, y, z).as_vec3() + 0.5) * scale - 0.5;
                    let pos = global_transform.transform_point(uvw);
                    let voxel_cube = commands
                        .spawn((
                            Mesh3d(example_assets.voxel_cube.clone()),
                            MeshMaterial3d(voxel_cube_material.clone()),
                            Transform::from_scale(Vec3::splat(VOXEL_CUBE_SCALE))
                                .with_translation(pos),
                        ))
                        .insert(VoxelCube)
                        .insert(NotShadowCaster)
                        .id();

                    commands.entity(voxel_cube_parent).add_child(voxel_cube);
                }
            }
        }
    }
}

pub fn try_inverse(&self) -> Option<Mat4>

Returns the inverse of self or None if the matrix is not invertible, or if the inverse is not finite.

pub fn inverse_or_zero(&self) -> Mat4

Returns the inverse of self or Mat4::ZERO if the matrix is not invertible, or if the inverse is not finite.

pub fn look_to_lh(eye: Vec3, dir: Vec3, up: Vec3) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::lh::view::look_to_mat4 function instead

Creates a left-handed view matrix using a camera position, a facing direction and an up direction

For a view coordinate system with +X=right, +Y=up and +Z=forward.

Panics

Will panic if dir or up are not normalized, or if dir and up are parallel, when glam_assert is enabled.

pub fn look_to_rh(eye: Vec3, dir: Vec3, up: Vec3) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::view::look_to_mat4 function instead

Creates a right-handed view matrix using a camera position, a facing direction, and an up direction.

For a view coordinate system with +X=right, +Y=up and +Z=back.

Panics

Will panic if dir or up are not normalized, or if dir and up are parallel, when glam_assert is enabled.

pub fn look_at_lh(eye: Vec3, center: Vec3, up: Vec3) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::lh::view::look_at_mat4 function instead

Creates a left-handed view matrix using a camera position, a focal points and an up direction.

For a view coordinate system with +X=right, +Y=up and +Z=forward.

Panics

Will panic if up is not normalized, if center is equal to eye, or if the view direction is parallel to up, when glam_assert is enabled.

pub fn look_at_rh(eye: Vec3, center: Vec3, up: Vec3) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::view::look_at_mat4 function instead

Creates a right-handed view matrix using a camera position, a focal point, and an up direction.

For a view coordinate system with +X=right, +Y=up and +Z=back.

Panics

Will panic if up is not normalized, if center is equal to eye, or if the view direction is parallel to up, when glam_assert is enabled.

pub fn frustum_rh_gl( left: f32, right: f32, bottom: f32, top: f32, z_near: f32, z_far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::proj::opengl::frustum function instead

Creates a right-handed perspective projection matrix with [-1,1] depth range.

This is the same as the OpenGL glFrustum function.

See https://registry.khronos.org/OpenGL-Refpages/gl2.1/xhtml/glFrustum.xml

pub fn frustum_lh( left: f32, right: f32, bottom: f32, top: f32, z_near: f32, z_far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::lh::proj::directx::frustum function instead

Creates a left-handed perspective projection matrix with [0,1] depth range.

Panics

Will panic if left equals right, bottom equals top, z_near equals z_far, or z_near or z_far are not positive when glam_assert is enabled.

pub fn frustum_rh( left: f32, right: f32, bottom: f32, top: f32, z_near: f32, z_far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::proj::directx::frustum function instead

Creates a right-handed perspective projection matrix with [0,1] depth range.

Panics

Will panic if left equals right, bottom equals top, z_near equals z_far, or z_near or z_far are not positive when glam_assert is enabled.

pub fn perspective_rh_gl( fov_y_radians: f32, aspect_ratio: f32, z_near: f32, z_far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::proj::opengl::perspective function instead

Creates a right-handed perspective projection matrix with [-1,1] depth range.

Useful to map the standard right-handed coordinate system into what OpenGL expects.

This is the same as the OpenGL gluPerspective function. See https://www.khronos.org/registry/OpenGL-Refpages/gl2.1/xhtml/gluPerspective.xml

pub fn perspective_lh( fov_y_radians: f32, aspect_ratio: f32, z_near: f32, z_far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::lh::proj::directx::perspective function instead

Creates a left-handed perspective projection matrix with [0,1] depth range.

Useful to map the standard left-handed coordinate system into what WebGPU/Metal/Direct3D expect.

Panics

Will panic if fov_y_radians is not in the range (0, π), if aspect_ratio is zero, or if z_near or z_far are less than or equal to zero, or if z_near is equal to z_far, when glam_assert is enabled.

pub fn perspective_rh( fov_y_radians: f32, aspect_ratio: f32, z_near: f32, z_far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::proj::directx::perspective function instead

Creates a right-handed perspective projection matrix with [0,1] depth range.

Useful to map the standard right-handed coordinate system into what WebGPU/Metal/Direct3D expect.

Panics

Will panic if fov_y_radians is not in the range (0, π), if aspect_ratio is zero, or if z_near or z_far are less than or equal to zero, or if z_near is equal to z_far, when glam_assert is enabled.

pub fn perspective_infinite_lh( fov_y_radians: f32, aspect_ratio: f32, z_near: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::lh::proj::directx::perspective_infinite function instead

Creates an infinite left-handed perspective projection matrix with [0,1] depth range.

Like perspective_lh, but with an infinite value for z_far. The result is that points near z_near are mapped to depth 0, and as they move towards infinity the depth approaches 1.

Panics

Will panic if fov_y_radians is not in the range (0, π), if aspect_ratio is zero, or if z_near is less than or equal to zero when glam_assert is enabled.

pub fn perspective_infinite_reverse_lh( fov_y_radians: f32, aspect_ratio: f32, z_near: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::lh::proj::directx::perspective_infinite_reverse function instead

Creates an infinite reverse left-handed perspective projection matrix with [0,1] depth range.

Similar to perspective_infinite_lh, but maps Z = z_near to a depth of 1 and Z = infinity to a depth of 0.

Panics

Will panic if fov_y_radians is not in the range (0, π), if aspect_ratio is zero, or if z_near is less than or equal to zero when glam_assert is enabled.

pub fn perspective_infinite_rh( fov_y_radians: f32, aspect_ratio: f32, z_near: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::proj::directx::perspective_infinite function instead

Creates an infinite right-handed perspective projection matrix with [0,1] depth range.

Like perspective_rh, but with an infinite value for z_far. The result is that points near z_near are mapped to depth 0, and as they move towards infinity the depth approaches 1.

Panics

Will panic if fov_y_radians is not in the range (0, π), if aspect_ratio is zero, or if z_near is less than or equal to zero when glam_assert is enabled.

pub fn perspective_infinite_reverse_rh( fov_y_radians: f32, aspect_ratio: f32, z_near: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::proj::directx::perspective_infinite_reverse function instead

Creates an infinite reverse right-handed perspective projection matrix with [0,1] depth range.

Similar to perspective_infinite_rh, but maps Z = z_near to a depth of 1 and Z = infinity to a depth of 0.

Panics

Will panic if fov_y_radians is not in the range (0, π), if aspect_ratio is zero, or if z_near is less than or equal to zero when glam_assert is enabled.

pub fn orthographic_rh_gl( left: f32, right: f32, bottom: f32, top: f32, near: f32, far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::proj::opengl::orthographic function instead

Creates a right-handed orthographic projection matrix with [-1,1] depth range. This is the same as the OpenGL glOrtho function in OpenGL. See https://www.khronos.org/registry/OpenGL-Refpages/gl2.1/xhtml/glOrtho.xml

Useful to map a right-handed coordinate system to the normalized device coordinates that OpenGL expects.

pub fn orthographic_lh( left: f32, right: f32, bottom: f32, top: f32, near: f32, far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::lh::proj::directx::orthographic function instead

Creates a left-handed orthographic projection matrix with [0,1] depth range.

Useful to map a left-handed coordinate system to the normalized device coordinates that WebGPU/Direct3D/Metal expect.

pub fn orthographic_rh( left: f32, right: f32, bottom: f32, top: f32, near: f32, far: f32, ) -> Mat4

👎Deprecated since 0.33.1:

use the glam::camera::rh::proj::directx::orthographic function instead

Creates a right-handed orthographic projection matrix with [0,1] depth range.

Useful to map a right-handed coordinate system to the normalized device coordinates that WebGPU/Direct3D/Metal expect.

pub fn project_point3(&self, rhs: Vec3) -> Vec3

Transforms the given 3D vector as a point, applying perspective correction.

This is the equivalent of multiplying the 3D vector as a 4D vector where w is 1.0. The perspective divide is performed meaning the resulting 3D vector is divided by w.

This method assumes that self contains a projective transform.

pub fn transform_point3(&self, rhs: Vec3) -> Vec3

Transforms the given 3D vector as a point.

This is the equivalent of multiplying the 3D vector as a 4D vector where w is 1.0.

This method assumes that self contains a valid affine transform. It does not perform a perspective divide, if self contains a perspective transform, or if you are unsure, the Self::project_point3() method should be used instead.

Panics

Will panic if the 3rd row of self is not (0, 0, 0, 1) when glam_assert is enabled.

pub fn transform_vector3(&self, rhs: Vec3) -> Vec3

Transforms the given 3D vector as a direction.

This is the equivalent of multiplying the 3D vector as a 4D vector where w is 0.0.

This method assumes that self contains a valid affine transform.

Panics

Will panic if the 3rd row of self is not (0, 0, 0, 1) when glam_assert is enabled.

pub fn project_point3a(&self, rhs: Vec3A) -> Vec3A

Transforms the given Vec3A as a 3D point, applying perspective correction.

This is the equivalent of multiplying the Vec3A as a 4D vector where w is 1.0. The perspective divide is performed meaning the resulting 3D vector is divided by w.

This method assumes that self contains a projective transform.

pub fn transform_point3a(&self, rhs: Vec3A) -> Vec3A

Transforms the given Vec3A as 3D point.

This is the equivalent of multiplying the Vec3A as a 4D vector where w is 1.0.

This method assumes that self contains a valid affine transform. If self contains a perspective transform, or if you are unsure, the Self::project_point3a() method should be used instead.

Panics

Will panic if the 3rd row of self is not (0, 0, 0, 1) when glam_assert is enabled.

pub fn transform_vector3a(&self, rhs: Vec3A) -> Vec3A

Transforms the given Vec3A as 3D vector.

This is the equivalent of multiplying the Vec3A as a 4D vector where w is 0.0.

This method assumes that self contains a valid affine transform.

Panics

Will panic if the 3rd row of self is not (0, 0, 0, 1) when glam_assert is enabled.

pub fn mul_vec4(&self, rhs: Vec4) -> Vec4

Transforms a 4D vector.

pub fn mul_transpose_vec4(&self, rhs: Vec4) -> Vec4

Transforms a 4D vector by the transpose of self.

pub fn mul_mat4(&self, rhs: &Mat4) -> Mat4

Multiplies two 4x4 matrices.

pub fn add_mat4(&self, rhs: &Mat4) -> Mat4

Adds two 4x4 matrices.

pub fn sub_mat4(&self, rhs: &Mat4) -> Mat4

Subtracts two 4x4 matrices.

pub fn mul_scalar(&self, rhs: f32) -> Mat4

Multiplies a 4x4 matrix by a scalar.

pub fn mul_diagonal_scale(&self, scale: Vec4) -> Mat4

Multiply self by a scaling vector scale. This is faster than creating a whole diagonal scaling matrix and then multiplying that. This operation is commutative.

pub fn div_scalar(&self, rhs: f32) -> Mat4

Divides a 4x4 matrix by a scalar.

pub fn recip(&self) -> Mat4

Returns a matrix containing the reciprocal 1.0/n of each element of self.

pub fn abs_diff_eq(&self, rhs: Mat4, max_abs_diff: f32) -> bool

Returns true if the absolute difference of all elements between self and rhs is less than or equal to max_abs_diff.

This can be used to compare if two matrices contain similar elements. It works best when comparing with a known value. The max_abs_diff that should be used used depends on the values being compared against.

For more see comparing floating point numbers.

pub fn abs(&self) -> Mat4

Takes the absolute value of each element in self

pub fn as_dmat4(&self) -> DMat4

Available on crate feature f64 only.

Examples found in repository

examples/camera/pan_orbit_camera_custom_input_plugin.rs (line 219)

    pub fn receive_messages(
        mut events: MessageReader<Self>,
        mut controllers: Query<(&mut PanOrbitCamera, &GlobalTransform)>,
        mut camera_map: ResMut<CameraPointerMap>,
        pointer_map: Res<PointerMap>,
        pointer_interactions: Query<&PointerInteraction>,
        pointer_locations: Query<&PointerLocation>,
        cameras: Query<(&Camera, &Projection)>,
        windows: Query<&Window>,
    ) {
        for event in events.read() {
            let Ok((mut controller, cam_transform)) = controllers.get_mut(event.camera()) else {
                continue;
            };

            match event {
                PanOrbitCameraInputMessage::Start { kind, pointer, .. } => {
                    if controller.is_actively_controlled() {
                        continue;
                    }
                    let anchor = pointer_map
                        .get_entity(*pointer)
                        .and_then(|entity| pointer_interactions.get(entity).ok())
                        .and_then(|interaction| interaction.get_nearest_hit())
                        // Since `bevy` 0.17.3:
                        //
                        // If the current hit is on a window, we cannot use the `hit.position` as an anchor
                        // as the `hit.position` is in viewport coordinates.
                        .filter(|(entity, _hit)| !windows.contains(*entity))
                        .and_then(|(_, hit)| hit.position)
                        .map(|world_space_hit| {
                            // Convert the world space hit to view (camera) space
                            cam_transform
                                .to_matrix()
                                .as_dmat4()
                                .inverse()
                                .transform_point3(world_space_hit.into())
                        })
                        .filter(|p| {
                            #[cfg(debug_assertions)]
                            if !p.is_finite() {
                                bevy::log::warn!(
                                    "Non-finite input fed to camera controller: {p:?}"
                                );
                            }
                            p.is_finite()
                        })
                        .or_else(|| {
                            let camera = cameras.get(event.camera()).ok();
                            let pointer_location = pointer_map
                                .get_entity(*pointer)
                                .and_then(|entity| pointer_locations.get(entity).ok())
                                .and_then(|l| l.location());
                            if let Some(((camera, proj), pointer_location)) =
                                camera.zip(pointer_location)
                            {
                                screen_to_view_space(
                                    camera,
                                    proj,
                                    &controller,
                                    pointer_location.position,
                                )
                            } else {
                                None
                            }
                        })
                        .filter(|p| p.is_finite());

                    match kind {
                        MotionKind::OrbitZoom => controller.start_orbit(anchor),
                        MotionKind::PanZoom => controller.start_pan(anchor),
                        MotionKind::Zoom => controller.start_zoom(anchor),
                    }
                    camera_map.insert(*pointer, event.camera());
                }
                PanOrbitCameraInputMessage::End { .. } => {
                    controller.end_move();
                    if let Some(pointer) = camera_map
                        .iter()
                        .find(|(.., camera)| **camera == event.camera())
                        .map(|(&pointer, ..)| pointer)
                    {
                        camera_map.remove(&pointer);
                    }
                }
            }
        }
    }

    /// While a camera motion is active, this system will take care of sending new pointer motion to
    /// the camera controller. The camera controller assumes that pan and orbit movements are tied
    /// to screen space pointer motion.
    ///
    /// This is because some of the pixel-perfect features of the controller require that data be
    /// passed in as screen space deltas, to compute perfect first-order control. This is also
    /// because the plugin uses pointer information to know which camera is being controlled.
    ///
    /// If you want to control the camera with different inputs, you will need to replace this
    /// system with one that tracks other input methods, and sends the required zoom and screenspace
    /// movement information.
    pub fn send_pointer_inputs(
        camera_map: Res<CameraPointerMap>,
        mut camera_controllers: Query<&mut PanOrbitCamera>,
        mut mouse_wheel: MessageReader<MouseWheel>,
        mut moves: MessageReader<PointerInput>,
    ) {
        let moves_list: Vec<_> = moves.read().collect();
        for (pointer, camera) in camera_map.iter() {
            let Ok(mut camera_controller) = camera_controllers.get_mut(*camera) else {
                continue;
            };

            let screenspace_input = moves_list
                .iter()
                .filter(|m| m.pointer_id.eq(pointer))
                .filter_map(|m| match m.action {
                    PointerAction::Move { delta } => Some(delta),
                    _ => None,
                })
                .sum();

            let zoom_amount = match pointer {
                // TODO: add pinch zoom support
                PointerId::Mouse => mouse_wheel
                    .read()
                    .map(|mw| {
                        let scroll_multiplier = match mw.unit {
                            MouseScrollUnit::Line => 150.0,
                            MouseScrollUnit::Pixel => 1.0,
                        };
                        mw.y * scroll_multiplier
                    })
                    .sum::<f32>(),
                _ => 0.0,
            };

            camera_controller.send_screenspace_input(screenspace_input);
            camera_controller.send_zoom_input(zoom_amount);
        }
        // This must be cleared manually because reading these inputs is conditional - we are not
        // guaranteed to be flushing the events every frame.
        mouse_wheel.clear();
    }
}

fn screen_to_view_space(
    camera: &Camera,
    proj: &Projection,
    controller: &PanOrbitCamera,
    target_position: Vec2,
) -> Option<DVec3> {
    let mut viewport_position = if let Some(rect) = camera.logical_viewport_rect() {
        target_position.as_dvec2() - rect.min.as_dvec2()
    } else {
        target_position.as_dvec2()
    };
    let target_size = camera.logical_viewport_size()?.as_dvec2();
    // Flip the Y co-ordinate origin from the top to the bottom.
    viewport_position.y = target_size.y - viewport_position.y;
    let ndc = viewport_position * 2. / target_size - DVec2::ONE;
    let ndc_to_view = proj.get_clip_from_view().as_dmat4().inverse();
    let view_near_plane = ndc_to_view.project_point3(ndc.extend(1.));
    match &proj {
        Projection::Perspective(_) | Projection::Custom(_) => {
            // Using EPSILON because an NDC with Z = 0 returns NaNs.
            let view_far_plane = ndc_to_view.project_point3(ndc.extend(f64::EPSILON));
            let direction = (view_far_plane - view_near_plane).normalize();
            Some((direction / direction.z) * controller.last_anchor_depth())
        }
        Projection::Orthographic(_) => Some(DVec3::new(
            view_near_plane.x,
            view_near_plane.y,
            controller.last_anchor_depth(),
        )),
    }
}

Trait Implementations

impl AbsDiffEq for Mat4

type Epsilon = <f32 as AbsDiffEq>::Epsilon

Used for specifying relative comparisons.

fn default_epsilon() -> <Mat4 as AbsDiffEq>::Epsilon

The default tolerance to use when testing values that are close together.

fn abs_diff_eq( &self, other: &Mat4, epsilon: <Mat4 as AbsDiffEq>::Epsilon, ) -> bool

A test for equality that uses the absolute difference to compute the approximate equality of two numbers.

fn abs_diff_ne(&self, other: &Rhs, epsilon: Self::Epsilon) -> bool

The inverse of AbsDiffEq::abs_diff_eq.

impl Add for Mat4

type Output = Mat4

The resulting type after applying the + operator.

fn add(self, rhs: Mat4) -> Mat4

Performs the + operation.

impl Add<&Mat4> for Mat4

type Output = Mat4

The resulting type after applying the + operator.

fn add(self, rhs: &Mat4) -> Mat4

Performs the + operation.

impl Add<&Mat4> for &Mat4

type Output = Mat4

The resulting type after applying the + operator.

fn add(self, rhs: &Mat4) -> Mat4

Performs the + operation.

impl Add<Mat4> for &Mat4

type Output = Mat4

The resulting type after applying the + operator.

fn add(self, rhs: Mat4) -> Mat4

Performs the + operation.

impl AddAssign for Mat4

fn add_assign(&mut self, rhs: Mat4)

Performs the += operation.

impl AddAssign<&Mat4> for Mat4

fn add_assign(&mut self, rhs: &Mat4)

Performs the += operation.

impl AsMut<[f32; 16]> for Mat4

fn as_mut(&mut self) -> &mut [f32; 16]

Converts this type into a mutable reference of the (usually inferred) input type.

impl AsMutMatrixParts<f32, 4, 4> for Mat4

where Mat4: AsMut<[f32; 16]>, f32: MatrixScalar,

fn as_mut_parts(&mut self) -> &mut [[f32; 4]; 4]

impl AsRef<[f32; 16]> for Mat4

fn as_ref(&self) -> &[f32; 16]

Converts this type into a shared reference of the (usually inferred) input type.

impl AsRefMatrixParts<f32, 4, 4> for Mat4

where Mat4: AsRef<[f32; 16]>, f32: MatrixScalar,

fn as_ref_parts(&self) -> &[[f32; 4]; 4]

impl Clone for Mat4

fn clone(&self) -> Mat4

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for Mat4

impl CreateFrom for Mat4

where Mat4: ShaderType<ExtraMetadata = MatrixMetadata> + FromMatrixParts<f32, 4, 4>, f32: MatrixScalar + CreateFrom,

fn create_from<B>(reader: &mut Reader<B>) -> Mat4

where B: BufferRef,

impl Debug for Mat4

fn fmt(&self, fmt: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Default for Mat4

fn default() -> Mat4

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for Mat4

Deserialize expects a sequence of 16 values.

fn deserialize<D>( deserializer: D, ) -> Result<Mat4, <D as Deserializer<'de>>::Error>

where D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Display for Mat4

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.

impl Div<&f32> for Mat4

type Output = Mat4

The resulting type after applying the / operator.

fn div(self, rhs: &f32) -> Mat4

Performs the / operation.

impl Div<&f32> for &Mat4

type Output = Mat4

The resulting type after applying the / operator.

fn div(self, rhs: &f32) -> Mat4

Performs the / operation.

impl Div<f32> for Mat4

type Output = Mat4

The resulting type after applying the / operator.

fn div(self, rhs: f32) -> Mat4

Performs the / operation.

impl Div<f32> for &Mat4

type Output = Mat4

The resulting type after applying the / operator.

fn div(self, rhs: f32) -> Mat4

Performs the / operation.

impl DivAssign<&f32> for Mat4

fn div_assign(&mut self, rhs: &f32)

Performs the /= operation.

impl DivAssign<f32> for Mat4

fn div_assign(&mut self, rhs: f32)

Performs the /= operation.

impl From<Affine3> for Mat4

fn from(m: Affine3) -> Mat4

Converts to this type from the input type.

impl From<Affine3A> for Mat4

fn from(m: Affine3A) -> Mat4

Converts to this type from the input type.

impl From<Mat4> for GlobalTransform

fn from(world_from_local: Mat4) -> GlobalTransform

Converts to this type from the input type.

impl FromArg for Mat4

type This<'from_arg> = Mat4

The type to convert into.

fn from_arg(arg: Arg<'_>) -> Result<<Mat4 as FromArg>::This<'_>, ArgError>

Creates an item from an argument.

impl FromMatrixParts<f32, 4, 4> for Mat4

fn from_parts(parts: [[f32; 4]; 4]) -> Mat4

impl FromReflect for Mat4

fn from_reflect(reflect: &(dyn PartialReflect + 'static)) -> Option<Mat4>

Constructs a concrete instance of Self from a reflected value.

fn take_from_reflect( reflect: Box<dyn PartialReflect>, ) -> Result<Self, Box<dyn PartialReflect>>

Attempts to downcast the given value to Self using, constructing the value using from_reflect if that fails.

impl GetOwnership for Mat4

fn ownership() -> Ownership

Returns the ownership of Self.

impl GetTypeRegistration for Mat4

fn get_type_registration() -> TypeRegistration

Returns the default TypeRegistration for this type.

fn register_type_dependencies(registry: &mut TypeRegistry)

Registers other types needed by this type.

impl IntoReturn for Mat4

fn into_return<'into_return>(self) -> Return<'into_return>

where Mat4: 'into_return,

Converts Self into a Return value.

impl Mul for Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Mat4) -> Mat4

Performs the * operation.

impl Mul<&Affine3> for Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Affine3) -> Mat4

Performs the * operation.

impl Mul<&Affine3> for &Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Affine3) -> Mat4

Performs the * operation.

impl Mul<&Affine3A> for Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Affine3A) -> Mat4

Performs the * operation.

impl Mul<&Affine3A> for &Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Affine3A) -> Mat4

Performs the * operation.

impl Mul<&Mat4> for Affine3

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Mat4) -> Mat4

Performs the * operation.

impl Mul<&Mat4> for &Affine3

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Mat4) -> Mat4

Performs the * operation.

impl Mul<&Mat4> for Affine3A

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Mat4) -> Mat4

Performs the * operation.

impl Mul<&Mat4> for &Affine3A

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Mat4) -> Mat4

Performs the * operation.

impl Mul<&Mat4> for Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Mat4) -> Mat4

Performs the * operation.

impl Mul<&Mat4> for &Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &Mat4) -> Mat4

Performs the * operation.

impl Mul<&Vec4> for Mat4

type Output = Vec4

The resulting type after applying the * operator.

fn mul(self, rhs: &Vec4) -> Vec4

Performs the * operation.

impl Mul<&Vec4> for &Mat4

type Output = Vec4

The resulting type after applying the * operator.

fn mul(self, rhs: &Vec4) -> Vec4

Performs the * operation.

impl Mul<&f32> for Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &f32) -> Mat4

Performs the * operation.

impl Mul<&f32> for &Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: &f32) -> Mat4

Performs the * operation.

impl Mul<Affine3> for Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Affine3) -> Mat4

Performs the * operation.

impl Mul<Affine3> for &Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Affine3) -> Mat4

Performs the * operation.

impl Mul<Affine3A> for Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Affine3A) -> Mat4

Performs the * operation.

impl Mul<Affine3A> for &Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Affine3A) -> Mat4

Performs the * operation.

impl Mul<Mat4> for Affine3

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Mat4) -> <Affine3 as Mul<Mat4>>::Output

Performs the * operation.

impl Mul<Mat4> for &Affine3

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Mat4) -> Mat4

Performs the * operation.

impl Mul<Mat4> for Affine3A

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Mat4) -> <Affine3A as Mul<Mat4>>::Output

Performs the * operation.

impl Mul<Mat4> for &Affine3A

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Mat4) -> Mat4

Performs the * operation.

impl Mul<Mat4> for &Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: Mat4) -> Mat4

Performs the * operation.

impl Mul<Vec4> for Mat4

type Output = Vec4

The resulting type after applying the * operator.

fn mul(self, rhs: Vec4) -> <Mat4 as Mul<Vec4>>::Output

Performs the * operation.

impl Mul<Vec4> for &Mat4

type Output = Vec4

The resulting type after applying the * operator.

fn mul(self, rhs: Vec4) -> Vec4

Performs the * operation.

impl Mul<f32> for Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: f32) -> Mat4

Performs the * operation.

impl Mul<f32> for &Mat4

type Output = Mat4

The resulting type after applying the * operator.

fn mul(self, rhs: f32) -> Mat4

Performs the * operation.

impl MulAssign for Mat4

fn mul_assign(&mut self, rhs: Mat4)

Performs the *= operation.

impl MulAssign<&Affine3> for Mat4

fn mul_assign(&mut self, rhs: &Affine3)

Performs the *= operation.

impl MulAssign<&Affine3A> for Mat4

fn mul_assign(&mut self, rhs: &Affine3A)

Performs the *= operation.

impl MulAssign<&Mat4> for Mat4

fn mul_assign(&mut self, rhs: &Mat4)

Performs the *= operation.

impl MulAssign<&f32> for Mat4

fn mul_assign(&mut self, rhs: &f32)

Performs the *= operation.

impl MulAssign<Affine3> for Mat4

fn mul_assign(&mut self, rhs: Affine3)

Performs the *= operation.

impl MulAssign<Affine3A> for Mat4

fn mul_assign(&mut self, rhs: Affine3A)

Performs the *= operation.

impl MulAssign<f32> for Mat4

fn mul_assign(&mut self, rhs: f32)

Performs the *= operation.

impl Neg for Mat4

type Output = Mat4

The resulting type after applying the - operator.

fn neg(self) -> <Mat4 as Neg>::Output

Performs the unary - operation.

impl Neg for &Mat4

type Output = Mat4

The resulting type after applying the - operator.

fn neg(self) -> Mat4

Performs the unary - operation.

impl PartialEq for Mat4

fn eq(&self, rhs: &Mat4) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl PartialReflect for Mat4

fn get_represented_type_info(&self) -> Option<&'static TypeInfo>

Returns the TypeInfo of the type represented by this value.

fn try_apply( &mut self, value: &(dyn PartialReflect + 'static), ) -> Result<(), ApplyError>

Tries to apply a reflected value to this value.

fn reflect_kind(&self) -> ReflectKind

Returns a zero-sized enumeration of “kinds” of type.

fn reflect_ref(&self) -> ReflectRef<'_>

Returns an immutable enumeration of “kinds” of type.

fn reflect_mut(&mut self) -> ReflectMut<'_>

Returns a mutable enumeration of “kinds” of type.

fn reflect_owned(self: Box<Mat4>) -> ReflectOwned

Returns an owned enumeration of “kinds” of type.

fn try_into_reflect( self: Box<Mat4>, ) -> Result<Box<dyn Reflect>, Box<dyn PartialReflect>>

Attempts to cast this type to a boxed, fully-reflected value.

fn try_as_reflect(&self) -> Option<&(dyn Reflect + 'static)>

Attempts to cast this type to a fully-reflected value.

fn try_as_reflect_mut(&mut self) -> Option<&mut (dyn Reflect + 'static)>

Attempts to cast this type to a mutable, fully-reflected value.

fn into_partial_reflect(self: Box<Mat4>) -> Box<dyn PartialReflect>

Casts this type to a boxed, reflected value.

fn as_partial_reflect(&self) -> &(dyn PartialReflect + 'static)

Casts this type to a reflected value.

fn as_partial_reflect_mut(&mut self) -> &mut (dyn PartialReflect + 'static)

Casts this type to a mutable, reflected value.

fn reflect_partial_eq( &self, value: &(dyn PartialReflect + 'static), ) -> Option<bool>

Returns a “partial equality” comparison result.

fn reflect_partial_cmp( &self, value: &(dyn PartialReflect + 'static), ) -> Option<Ordering>

Returns a “partial comparison” result.

fn debug(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Debug formatter for the value.

fn reflect_clone(&self) -> Result<Box<dyn Reflect>, ReflectCloneError>

Attempts to clone Self using reflection.

fn apply(&mut self, value: &(dyn PartialReflect + 'static))

Applies a reflected value to this value.

fn to_dynamic(&self) -> Result<Box<dyn PartialReflect>, ReflectCloneError>

Converts this reflected value into its dynamic representation based on its kind.

fn reflect_clone_and_take<T>(&self) -> Result<T, ReflectCloneError>

where T: 'static, Self: Sized + TypePath,

For a type implementing PartialReflect, combines reflect_clone and take in a useful fashion, automatically constructing an appropriate ReflectCloneError if the downcast fails.

fn reflect_hash(&self) -> Option<u64>

Returns a hash of the value (which includes the type).

fn is_dynamic(&self) -> bool

Indicates whether or not this type is a dynamic type.

impl Pod for Mat4

impl Product for Mat4

fn product<I>(iter: I) -> Mat4

where I: Iterator<Item = Mat4>,

Takes an iterator and generates Self from the elements by multiplying the items.

impl<'a> Product<&'a Mat4> for Mat4

fn product<I>(iter: I) -> Mat4

where I: Iterator<Item = &'a Mat4>,

Takes an iterator and generates Self from the elements by multiplying the items.

impl ReadFrom for Mat4

where Mat4: ShaderType<ExtraMetadata = MatrixMetadata> + AsMutMatrixParts<f32, 4, 4>, f32: MatrixScalar + ReadFrom,

fn read_from<B>(&mut self, reader: &mut Reader<B>)

where B: BufferRef,

impl Reflect for Mat4

fn into_any(self: Box<Mat4>) -> Box<dyn Any>

Returns the value as a Box<dyn Any>.

fn as_any(&self) -> &(dyn Any + 'static)

Returns the value as a &dyn Any.

fn as_any_mut(&mut self) -> &mut (dyn Any + 'static)

Returns the value as a &mut dyn Any.

fn into_reflect(self: Box<Mat4>) -> Box<dyn Reflect>

Casts this type to a boxed, fully-reflected value.

fn as_reflect(&self) -> &(dyn Reflect + 'static)

Casts this type to a fully-reflected value.

fn as_reflect_mut(&mut self) -> &mut (dyn Reflect + 'static)

Casts this type to a mutable, fully-reflected value.

fn set(&mut self, value: Box<dyn Reflect>) -> Result<(), Box<dyn Reflect>>

Performs a type-checked assignment of a reflected value to this value.

impl RelativeEq for Mat4

fn default_max_relative() -> <Mat4 as AbsDiffEq>::Epsilon

The default relative tolerance for testing values that are far-apart.

fn relative_eq( &self, other: &Mat4, epsilon: <Mat4 as AbsDiffEq>::Epsilon, max_relative: <Mat4 as AbsDiffEq>::Epsilon, ) -> bool

A test for equality that uses a relative comparison if the values are far apart.

fn relative_ne( &self, other: &Rhs, epsilon: Self::Epsilon, max_relative: Self::Epsilon, ) -> bool

The inverse of RelativeEq::relative_eq.

impl Serialize for Mat4

Serialize as a sequence of 16 values.

fn serialize<S>( &self, serializer: S, ) -> Result<<S as Serializer>::Ok, <S as Serializer>::Error>

where S: Serializer,

Serialize this value into the given Serde serializer.

impl ShaderSize for Mat4

where f32: ShaderSize,

const SHADER_SIZE: NonZero<u64> = _

Represents WGSL Size (equivalent to ShaderType::min_size)

impl ShaderType for Mat4

where f32: ShaderSize,

fn min_size() -> NonZero<u64>

Represents the minimum size of Self (equivalent to GPUBufferBindingLayout.minBindingSize)

fn size(&self) -> NonZero<u64>

Returns the size of Self at runtime

fn assert_uniform_compat()

Asserts that Self meets the requirements of the uniform address space restrictions on stored values and the uniform address space layout constraints

impl Struct for Mat4

fn field(&self, name: &str) -> Option<&(dyn PartialReflect + 'static)>

Gets a reference to the value of the field named name as a &dyn PartialReflect.

fn field_mut( &mut self, name: &str, ) -> Option<&mut (dyn PartialReflect + 'static)>

Gets a mutable reference to the value of the field named name as a &mut dyn PartialReflect.

fn field_at(&self, index: usize) -> Option<&(dyn PartialReflect + 'static)>

Gets a reference to the value of the field with index index as a &dyn PartialReflect.

fn field_at_mut( &mut self, index: usize, ) -> Option<&mut (dyn PartialReflect + 'static)>

Gets a mutable reference to the value of the field with index index as a &mut dyn PartialReflect.

fn name_at(&self, index: usize) -> Option<&str>

Gets the name of the field with index index.

fn index_of_name(&self, name: &str) -> Option<usize>

Gets the index of the field with the given name.

fn field_len(&self) -> usize

Returns the number of fields in the struct.

fn iter_fields(&self) -> FieldIter<'_>

Returns an iterator over the values of the reflectable fields for this struct.

fn to_dynamic_struct(&self) -> Result<DynamicStruct, ReflectCloneError>

Creates a new DynamicStruct from this struct.

fn get_represented_struct_info(&self) -> Option<&'static StructInfo>

Will return None if TypeInfo is not available.

impl Sub for Mat4

type Output = Mat4

The resulting type after applying the - operator.

fn sub(self, rhs: Mat4) -> Mat4

Performs the - operation.

impl Sub<&Mat4> for Mat4

type Output = Mat4

The resulting type after applying the - operator.

fn sub(self, rhs: &Mat4) -> Mat4

Performs the - operation.

impl Sub<&Mat4> for &Mat4

type Output = Mat4

The resulting type after applying the - operator.

fn sub(self, rhs: &Mat4) -> Mat4

Performs the - operation.

impl Sub<Mat4> for &Mat4

type Output = Mat4

The resulting type after applying the - operator.

fn sub(self, rhs: Mat4) -> Mat4

Performs the - operation.

impl SubAssign for Mat4

fn sub_assign(&mut self, rhs: Mat4)

Performs the -= operation.

impl SubAssign<&Mat4> for Mat4

fn sub_assign(&mut self, rhs: &Mat4)

Performs the -= operation.

impl Sum for Mat4

fn sum<I>(iter: I) -> Mat4

where I: Iterator<Item = Mat4>,

Takes an iterator and generates Self from the elements by “summing up” the items.

impl<'a> Sum<&'a Mat4> for Mat4

fn sum<I>(iter: I) -> Mat4

where I: Iterator<Item = &'a Mat4>,

Takes an iterator and generates Self from the elements by “summing up” the items.

impl TransformPoint for Mat4

fn transform_point(&self, point: impl Into<Vec3>) -> Vec3

Transform a point.

impl TypePath for Mat4

fn type_path() -> &'static str

Returns the fully qualified path of the underlying type.

fn short_type_path() -> &'static str

Returns a short, pretty-print enabled path to the type.

fn type_ident() -> Option<&'static str>

Returns the name of the type, or None if it is anonymous.

fn crate_name() -> Option<&'static str>

Returns the name of the crate the type is in, or None if it is anonymous.

fn module_path() -> Option<&'static str>

Returns the path to the module the type is in, or None if it is anonymous.

impl Typed for Mat4

fn type_info() -> &'static TypeInfo

Returns the compile-time info for the underlying type.

impl UlpsEq for Mat4

fn default_max_ulps() -> u32

The default ULPs to tolerate when testing values that are far-apart.

fn ulps_eq( &self, other: &Mat4, epsilon: <Mat4 as AbsDiffEq>::Epsilon, max_ulps: u32, ) -> bool

A test for equality that uses units in the last place (ULP) if the values are far apart.

fn ulps_ne(&self, other: &Rhs, epsilon: Self::Epsilon, max_ulps: u32) -> bool

The inverse of UlpsEq::ulps_eq.

impl WriteInto for Mat4

where Mat4: ShaderType<ExtraMetadata = MatrixMetadata> + AsRefMatrixParts<f32, 4, 4>, f32: MatrixScalar + WriteInto,

fn write_into<B>(&self, writer: &mut Writer<B>)

where B: BufferMut,

impl Zeroable for Mat4

fn zeroed() -> Self

Calls zeroed.