Scene#

SceneUI is the framework built on top and tightly integrated with omni.ui. It uses omni.ui inputs and basically supports everything omni.ui supports, like Python bindings, properties, callbacks, and async workflow.

SceneView is the omni.ui widget that renders all the SceneUI items. It can be a part of the omni.ui layout or an overlay of omni.ui interface. It’s the entry point of SceneUI.

Camera#

SceneView determines the position and configuration of the camera and has projection and view matrices.

Code Result

# Projection matrix
proj = [1.7, 0, 0, 0, 0, 3, 0, 0, 0, 0, -1, -1, 0, 0, -2, 0]

# Move camera
rotation = sc.Matrix44.get_rotation_matrix(30, 50, 0, True)
transl = sc.Matrix44.get_translation_matrix(0, 0, -6)
view = transl * rotation

scene_view = sc.SceneView(
    sc.CameraModel(proj, view),
    aspect_ratio_policy=sc.AspectRatioPolicy.PRESERVE_ASPECT_FIT,
    height=200
)

with scene_view.scene:
    # Edges of cube
    sc.Line([-1, -1, -1], [1, -1, -1])
    sc.Line([-1, 1, -1], [1, 1, -1])
    sc.Line([-1, -1, 1], [1, -1, 1])
    sc.Line([-1, 1, 1], [1, 1, 1])

    sc.Line([-1, -1, -1], [-1, 1, -1])
    sc.Line([1, -1, -1], [1, 1, -1])
    sc.Line([-1, -1, 1], [-1, 1, 1])
    sc.Line([1, -1, 1], [1, 1, 1])

    sc.Line([-1, -1, -1], [-1, -1, 1])
    sc.Line([-1, 1, -1], [-1, 1, 1])
    sc.Line([1, -1, -1], [1, -1, 1])
    sc.Line([1, 1, -1], [1, 1, 1])

Screen#

To get position and view matrices, SceneView queries the model component. It represents either the data transferred from the external backend or the data that the model holds. The user can reimplement the model to manage the user input or get the camera directly from the renderer or any other back end.

The SceneView model is required to return two float arrays, projection and view of size 16, which are the camera matrices.

ui.Screen is designed to simplify tracking the user input to control the camera position. ui.Screen represents the rectangle always placed in the front of the camera. It doesn’t produce any visible shape, but it interacts with the user input.

You can change the camera orientation in the example below by dragging the mouse cursor left and right.

Code Result

class CameraModel(sc.AbstractManipulatorModel):
    def __init__(self):
        super().__init__()
        self._angle = 0

    def append_angle(self, delta: float):
        self._angle += delta * 100
        # Inform SceneView that view matrix is changed
        self._item_changed("view")

    def get_as_floats(self, item):
        """Called by SceneView to get projection and view matrices"""
        if item == self.get_item("projection"):
            # Projection matrix
            return [1, 0, 0, 0, 0, 1, 0, 0, 0, 0, -1, -1, 0, 0, -2, 0]
        if item == self.get_item("view"):
            # Move camera
            rotation = sc.Matrix44.get_rotation_matrix(30, self._angle, 0, True)
            transl = sc.Matrix44.get_translation_matrix(0, 0, -8)
            view = transl * rotation
            return [view[i] for i in range(16)]

def on_mouse_dragged(sender):
    # Change the model's angle according to mouse x offset
    mouse_moved = sender.gesture_payload.mouse_moved[0]
    sender.scene_view.model.append_angle(mouse_moved)

with sc.SceneView(CameraModel(), height=200).scene:
    # Camera control
    sc.Screen(gesture=sc.DragGesture(on_changed_fn=on_mouse_dragged))

    # Edges of cube
    sc.Line([-1, -1, -1], [1, -1, -1])
    sc.Line([-1, 1, -1], [1, 1, -1])
    sc.Line([-1, -1, 1], [1, -1, 1])
    sc.Line([-1, 1, 1], [1, 1, 1])

    sc.Line([-1, -1, -1], [-1, 1, -1])
    sc.Line([1, -1, -1], [1, 1, -1])
    sc.Line([-1, -1, 1], [-1, 1, 1])
    sc.Line([1, -1, 1], [1, 1, 1])

    sc.Line([-1, -1, -1], [-1, -1, 1])
    sc.Line([-1, 1, -1], [-1, 1, 1])
    sc.Line([1, -1, -1], [1, -1, 1])
    sc.Line([1, 1, -1], [1, 1, 1])
On this page