3D

Murali's 3D support is built for explanatory video scenes, not for game-engine style simulation. Use it when depth helps the viewer understand a shape, object, flow, or spatial relationship.

The current 3D workflow covers three main needs:

• authored mathematical geometry, such as curves, surfaces, axes, and wireframes
• static video props loaded from local .glb or .gltf files with Prop3D
• camera framing, movement, and perspective projection for making depth readable

All 3D objects are still ordinary Murali tattvas. Once an object is added to a scene, it can use the same scene-level placement and timeline transforms as other elements:

use glam::{Vec3, vec3};
use murali::frontend::collection::primitives::prop3d::Prop3D;

let prop_id = scene.add_tattva(
    Prop3D::from_file("assets/props/apple.glb")?,
    vec3(0.0, 0.0, 0.0),
);

scene.set_scale(prop_id, Vec3::splat(0.8));

timeline
    .animate(prop_id)
    .at(1.0)
    .for_duration(1.2)
    .move_to(vec3(1.0, 0.4, 0.0))
    .spawn();

What Exists Today

Murali currently supports:

• 3D positions through Vec3
• perspective and orthographic cameras
• 3D axes
• parametric 3D curves
• parametric surfaces
• wireframe and textured surfaces
• static .glb and .gltf props through Prop3D
• object movement, rotation, scaling, and opacity through existing timeline animations

Prop3D loads static glTF geometry, base-color material factors, and common base-color textures. Prefer .glb when available because it travels as one file. Loose .gltf works when the referenced .bin and texture files stay beside it. Prop3D is best for video props such as apples, balls, books, animals, tools, robots, and other simple objects that support a scene.

Current Boundaries

Murali does not currently try to be a full 3D engine. In particular, Prop3D does not yet support:

• skeletal animation
• physics or collision
• advanced PBR lighting
• engine-specific materials
• interactive object picking

Those boundaries are intentional. For Murali, the main question is whether a 3D object helps the video explain something clearly.

Making 3D Read Well

Depth is easy to add and easy to misuse. A good 3D scene usually needs at least one of these:

• a camera angle that separates front, side, and depth
• face colors, textures, or wireframes that reveal the form
• motion that shows the object from more than one angle
• surrounding axes, labels, or text that explain why the object matters

The bundled demo-pyramid.glb uses different colors on different faces because Murali's current mesh renderer is mostly unlit. Real props with textures often read better, but faceted color is a useful fallback for simple demo objects.

Recommended Examples

• curves_3d Focus: one traced parametric curve and camera framing
• surfaces_3d Focus: a translucent shaded surface whose form is clarified by color and camera motion
• wireframe_surfaces Focus: wireframe curvature reading
• textured_surface Focus: UV texture mapping on a globe
• prop3d_glb Focus: loading a single-file .glb prop and animating it with ordinary scene transforms. The bundled demo asset is faceted so its 3D form remains visible before full lighting support exists.
• prop3d_gltf Focus: loading a loose .gltf apple prop with its sibling .bin file.
• model_inspector Focus: resolving, centering, framing, and continuously rotating an unfamiliar GLB/GLTF asset before it is used in a scene.

Good First Pick

Start with curves_3d for the lightest 3D introduction, then move to surfaces_3d.

What To Look For

• whether the camera helps the eye understand the shape
• whether the scene uses 3D because it adds clarity, not just because it can
• whether surface reveals are staged clearly enough to read

Run Locally

cargo run --example curves_3d

For a static 3D prop loaded from a local GLB asset:

cargo run --example prop3d_glb

For a static 3D prop loaded from a loose glTF asset:

cargo run --example prop3d_gltf

To inspect either format with automatic framing and interactive camera controls:

cargo run --example model_inspector -- demo-apple