This collection gathers all terms from every lesson, starting with the studio’s key vocabulary. The terms are written in English, as they appear in the documentation, making them easier to find. You can click on any term in any lesson to open the glossary directly on that word.

Asset — a ready-made element for a scene, such as coral or a drone.

Location — the setting, assembled from assets, for example, a reef.

Shot — a segment of film from cut to cut.

Sequence — several shots within one location.

Department — a part of the team responsible for its specific work and layer.

Layout — the department responsible for arranging elements and the camera.

Lookdev — work on the appearance of objects, materials, and textures.

Publish — the folder containing assets handed off to other teams. Working scenes are kept in activities.

Scaffold — an empty master scene tree, with each branch representing a department.

stage — the entire scene, a USD assembled from layers. It is built in Houdini under /stage.

layer — a USD file containing a part of the scene.

layer stack — a stack of layers within a scene, ordered by strength.

opinion — an attribute value recorded by one layer. In a conflict, the opinion from the stronger layer prevails.

strength — the layer’s strength. In the SubLayer node, the last file is stronger by default, and in the list of subLayers within a USD file, the first entry is considered strongest.

prim — an object in the scene, a node in the Scene Graph Tree. Do not confuse it with a SOP primitive.

scenegraph — the scene graph, the entire scene tree composed of prims, which is displayed by the Scene Graph Tree.

schema — the type of prim, such as Xform, Scope, Mesh, or Material. In nodes, it may be called Primitive Type. The prim’s attributes depend on it.

Xform — a prim that moves everything inside it, for example, the asset root.

Scope — simply a folder that cannot be moved. This is how /World and its main branches are structured.

Mesh — a prim with polygonal geometry; for the coral, this includes render and proxy.

Gprim — a prim with geometry, such as a Mesh. Scene Doctor detects a Gprim inside another Gprim.

def — a layer that creates a prim itself; thus, the geo file creates a cache.

over — a layer that only modifies a prim from another layer. Without that source layer, there is nothing to modify.

Layer Break — the Houdini node; a new layer begins here. The entire scene is visible, but the USD ROP only writes changes after this point.

specifier — a tag indicating what a layer does to a prim: def, over, or class. In Houdini nodes, this is the Primitive Specifier.

class — a prim template that does not render itself. Other prims adopt its characteristics through inheritance.

sublayer — the entire file is stacked in layers; its prims maintain their respective paths. This is how the SubLayer node operates.

reference — a link that places a prim from a file onto any path in the scene. If you update the file, the scene updates as well.

composition arc — a general term for methods of assembling a scene from various files.

LIVERPS — the order of strength for these methods, from strongest to weakest. Local (layers), Inherits, VariantSets, rElocates, References, Payloads, Specializes. Previously, it was written as LIVRPS.

payload — a reference that can be omitted, which makes the heavy scene load faster.

inherit — a prim takes its opinion from the class. We modified the class, and all descendants were updated.

variant — a selectable option for a prim, such as one of two coral shapes.

relocates — moves a prim to a new path; the old path is closed afterward.

specializes — similar to inherit, but weaker than all methods.

kind — the role of a prim in the tree; objects are identified and the scene is validated based on it.

assembly — a high-level kind of scene section, such as /World and asset branches.

group — a kind of sub-section within an assembly, for example, corals.

component — the kind of the asset itself. Inside, there are only prims without a kind or subcomponent.

subcomponent — the kind of an important part of an asset, such as a car door.

model hierarchy — the chain of assembly or group kinds above an asset. As long as it is intact, the asset is a model—a cohesive scene object.

attribute — a named value of a prim, such as a displacement or color. It can change per frame.

xformOp — the translation, rotation, and scale of a prim, each with its own attribute, such as xformOp: translate.

purpose — who needs the prim. With render, only the renderer sees it; with proxy, only the viewport sees it; with default, everything sees it.

primvar — a geometry attribute, similar to in SOPs. It is seen by the renderer and the shader.

visibility — an attribute that hides the prim and everything within it.

defaultPrim — the main prim of the file; it is referenced without a prim path. This is always set, especially if there are multiple roots.

flatten — combining all layers into one file where there are no sublayers or references left. This is how assets are written in a single file.

root layer — the main scene file from which USD begins to assemble it. In Shot, this is the root.

Material — a primary material that contains shaders. It is assigned to the geometry, not the shader.

Shader — a node inside the Material, such as MtlX Standard Surface.

material: binding — a reference to the material, a path to the Material recorded on the primitive. Meshes that do not have their own receive this binding.

inputs — external parameters of the material, which are adjusted using Edit Material Properties.

prototype — a source model used to create copies. All copies inherit its geometry and material.

instance — a copy of the prototype. It only has its own position, rotation, and scale.

PointInstancer — a single prim for all copies. It contains points, each referencing the prototype, position, rotation, and scale.

instanceable — a marker on a prim with a reference. This allows USD to store identical copies in memory only once.

instance proxy — a prim inside a copy marked as instanceable. It is visible in the scene graph but cannot be modified.

startTimeCode and endTimeCode—the first and last frame in the USD file; they are written in the root shot.

time samples—attribute values per frame; this is how animation is stored in the layer.

value clips—animation broken down into per-frame files. USD only reads the file for the specific frame.

manifest—a utility file alongside the value clips, listing the primitives and attributes from the frame files.

topology—a utility file containing everything in the value clips that does not change per frame.

Volume—the volume primitive, whose fields link to its specific frame’s VDB files.

collection — a list of primitives under a specific name, registered to a primitive, such as a lamp.

light linking — a rule defining which primitives a lamp illuminates. Without it, the lamp illuminates everything.

LPE — a string that describes the ray path from the camera to the light. Image channels are collected along this path.

RenderVar — one image channel, also known as an AOV, for example, beauty.

LPE Tag — a tag for a lamp used with LPE. The light from a lamp with this tag is output to its specific AOV, and how much of it to retain is determined in the compositing stage.