NCP-OUSD study plan

Define and distinguish prims, attributes, relationships, properties, and metadata, then write a one-page glossary from memory and correct it against the docs

Author a USD file by hand (no DCC) containing a hierarchy of Xform and Mesh prims with correct typeName, default prim, and upAxis metadata

Trace how a stage resolves a single reference and a single sublayer, writing out the LIVERPS strength ordering (Local, Inherits, VariantSets, References, Payloads, Specializes) with an example per letter

Build a three-layer stack (base, shot override, lighting) and predict which opinion wins for an overridden attribute before opening the stage to verify

Construct a variantSet with at least three variants and demonstrate switching variants changes composed geometry without editing the base layer

Implement a payload that can be loaded and unloaded, and record stage metrics (prim count, loaded payload count) before and after unloading

Demonstrate the difference between a reference and a payload on the same asset and explain when each is appropriate

Resolve a real or synthetic layer-stack conflict by choosing between a sublayer, a reference, and an inherit, justifying the choice in writing

Convert a repeated-geometry scene into a point instancer or class-based instanceable reference set and measure prim count and file size reduction

Audit an asset's prim hierarchy for unnecessary deep nesting and flatten or restructure it to reduce traversal cost

Identify which prims can be marked instanceable and which cannot, listing the constraints that block instancing

Measure load time and memory for the instanced versus non-instanced version of the same scene

Explain the Hydra architecture (front-end scene delegate, scene index, render delegate) and diagram the data flow from stage to rendered image

Write or adapt a minimal scene index observer that prints prim and attribute notices as they arrive

Register and use a custom schema (a simple typed schema with one attribute) via plugInfo and generated or hand-written schema classes

Apply a schema to prims and confirm the applied API and typed attribute appear correctly in the composed stage

Compare how materials and geometry survive a round trip through USD, glTF, and FBX, noting which data is lost or transformed in each direction

Export a USD scene to glTF and re-import it, then diff the prim hierarchy and attribute values against the original

Implement one schema validation rule (for example, required attribute presence or allowed value range) and run it against a sample asset

Document the units, up-axis, and coordinate convention handling required when exchanging assets between two pipelines

Build a small command-line tool that opens a stage, queries prims by type or metadata, and prints a report

Automate an asset publish step that validates, sets default prim and metadata, and writes a versioned output path

Integrate the tool into a scripted pipeline run (shell or Python entry point) with clear exit codes and error messages

Handle a failure path: make the tool report a missing reference or failed validation without crashing

Diagnose a deliberately broken layer stack (bad reference path, conflicting opinions, missing payload) using composition diagnostics and layer inspection tools

Use the stage's composition query facilities to explain why a given prim or attribute has its composed value

Reproduce a rendering discrepancy by inspecting the scene index and confirming which prims reach the render delegate

Write a triage checklist that maps common symptoms to likely composition or schema causes

Take a timed practice set across all eight domains and score it by domain to find remaining weak areas

Re-explain composition strength ordering, instancing rules, and Hydra data flow from memory without notes

Revisit every previously missed concept and write a corrected one-paragraph explanation for each

Assemble a personal cheat sheet of commands, APIs, and gotchas for pre-exam review