How can Bevy's documentation be improved?
Zooming an orthographic camera is surprisingly tricky; unlike ordinary cameras moving them further away from the object doesn't cause the size of the objects displayed to change as all their projection is parallel.
@TheRawMeatball provided this initial solution in #help:
fn zoom_system(
mut whl: EventReader<MouseWheel>,
mut cam: Query<(&mut Transform, &mut OrthographicProjection), With<MainCamera>>,
windows: Res<Windows>,
) {
let delta_zoom: f32 = whl.iter().map(|e| e.y).sum();
if delta_zoom == 0. {
return;
}
let (mut pos, mut cam) = cam.single_mut().unwrap();
let window = windows.get_primary().unwrap();
let window_size = Vec2::new(window.width(), window.height());
let mouse_normalized_screen_pos =
(window.cursor_position().unwrap() / window_size) * 2. - Vec2::ONE;
let mouse_world_pos = pos.translation.truncate()
+ mouse_normalized_screen_pos * Vec2::new(cam.right, cam.top) * cam.scale;
cam.scale -= ZOOM_SPEED * delta_zoom * cam.scale;
cam.scale = cam.scale.clamp(MIN_ZOOM, MAX_ZOOM);
pos.translation = (mouse_world_pos
- mouse_normalized_screen_pos * Vec2::new(cam.right, cam.top) * cam.scale)
.extend(pos.translation.z);
}This should be straightforward to verify and clean up. Adding panning or rotation might also be worth including in the same example.
How can Bevy's documentation be improved?
Zooming an orthographic camera is surprisingly tricky; unlike ordinary cameras moving them further away from the object doesn't cause the size of the objects displayed to change as all their projection is parallel.
@TheRawMeatball provided this initial solution in #help:
This should be straightforward to verify and clean up. Adding panning or rotation might also be worth including in the same example.