NCP-OUSD OpenUSD Development Questions and Answers

OpenUSD file format plugins belong under the Data Exchange topic because they expand how USD participates in interchange workflows. Their purpose is to allow OpenUSD to read, interpret, and in some cases write data using formats beyond the native USD file types such as .usd, .usda, .usdc, and .usdz. Through these plugins, external file formats can be exposed to USD as layers and can participate in composition arcs such as references, payloads, and sublayers. This allows pipelines to integrate heterogeneous asset sources while still using USD’s scene description model, composition engine, and layer-based workflows.

Option A is correct because it directly identifies the function of file format plugins: extending OpenUSD functionality for reading and writing various file formats. Option B is incorrect because visualization is only one possible downstream use of exchanged data, not the purpose of the plugin system. Option C is incorrect because translation between formats does not inherently guarantee lossless preservation of every schema, opinion, or semantic construct. Option D is incorrect because compression is not the primary role of file format plugins. This aligns with the NVIDIA OpenUSD Development Study Guide topic Data Exchange , especially file formats, connectors, and plugin-based interoperability.

Custom attributes in OpenUSD are user-defined properties authored on prims to store additional typed data beyond predefined schema attributes. NVIDIA’s Learn OpenUSD material states that custom attributes can hold “numeric values, strings, or arrays,” and that custom attributes are created with UsdPrim::CreateAttribute(), where Sdf.ValueTypeNames represents the attribute type. NVIDIA’s OpenUSD data type reference lists Asset with the USDA value type token asset, and StringArray with the USDA token string[], making asset paths and string arrays valid attribute value types.

Option B is correct because asset path values are represented through the asset value type and are commonly used for resource references such as textures or external asset identifiers. Option C is correct because string[] is a valid array value type. Option A is incorrect because USD does not have a native struct attribute type; NVIDIA recommends namespace-prefixed attributes for mapping grouped or compound fields. Option D is incorrect in this context because dictionaries are associated with metadata/customData patterns, not ordinary typed custom attributes. This aligns with Customizing USD → Custom Properties, Sdf.ValueTypeNames, Namespaced Attributes.

An opinion may not appear in the final composed stage because USD resolves competing opinions according to composition strength. NVIDIA’s Learn OpenUSD material explains that a layer stack is the ordered set of a layer and its recursively gathered sublayers, with the root layer considered strongest, followed by sublayers according to their composed order. It also states that strength ordering determines which opinion is composed into the final stage when multiple layer stacks or layers contribute data for the same prim or property. ( docs.nvidia.com )

Option D is correct because a weaker layer can author a valid opinion, but that opinion can be hidden by a stronger authored opinion on the same target. Option A is incorrect because sublayers do not need to be referenced by every other layer to affect a stage. Option B is incorrect because there is no general “weaker keyword” that lowers priority. Option C is incorrect because opinions from sublayers, references, payloads, variants, inherits, and specializes can all contribute to composition. This aligns with Composition → Sublayers, Layer Stacks, Opinion Strength, and LIVERPS Strength Ordering .

Codeless schemas are preferred when the pipeline goal is portability, easy distribution, and reduced application coupling. NVIDIA’s Learn OpenUSD schema guidance states that schemas define data models and optional APIs for encoding and interchanging 3D and non-3D concepts, and notes a trend toward codeless schemas for easier distribution, with schemas becoming more focused on data modeling rather than behavior implementation.

Option A is correct because a codeless schema can be distributed as schema data and plugin metadata without requiring every consuming application to compile, link, or ship custom generated C++/Python schema classes. OpenUSD’s schema-generation documentation identifies codeless schemas as schemas produced without corresponding C++ classes, where only generatedSchema.usda and plugInfo.json are essential for runtime registration.

Option B is incorrect because strongly typed convenience APIs are the advantage of codeful generated schemas. Option C is incorrect because fallback values are authored in schema definitions. Option D is incorrect because codeful and codeless schemas can both participate in schema registration and value interpretation. This aligns with Customizing USD → Schemas, API Schemas, Codeless Schemas, Schema Registry, Data Modeling for Interchange .

Option A is correct because local opinions are the strongest opinions in USD composition strength ordering. NVIDIA’s Learn OpenUSD strength-ordering guide describes LIVRPS, now also written as LIVERPS with relocates included, as the ordered set of composition operations used to determine which opinions win. The order is strongest to weakest: local , inherits, variant sets, relocates, references, payloads, and specializes.

When debugging an unexpected value, local opinions should be inspected first because a directly authored opinion in the local layer stack can override opinions arriving through other composition arcs. NVIDIA’s tracing guidance also highlights how complex composition debugging becomes when many prims, properties, layers, and arcs contribute competing opinions.

Option B is false because specializes arcs are not unoverrideable; they are the weakest fallback-style opinions in the composition order. Option C is false because references are stronger than payloads in LIVRPS/LIVERPS ordering, so checking payloads before references reverses the standard diagnostic priority. This aligns with Debugging and Troubleshooting → LIVERPS Strength Ordering, Local Opinions, References, Payloads, Specializes, and Composition Diagnostics .

The Python snippet authors a reference opinion directly on the prim /refSphere in a new anonymous root layer. Usd.Stage.CreateInMemory() creates an empty stage, and OverridePrim("/refSphere") authors an over prim spec because no concrete type is being defined. The call refSphere.GetReferences().AddReference("./HelloWorld.usda") adds a reference composition arc to that same prim, so the root layer serializes the opinion as prepend references = @./HelloWorld.usda@.

Option C is correct because ExportToString() prints the authored contents of the root layer, not the fully expanded composed result of the referenced asset. The referenced HelloWorld.usda layer has defaultPrim = "hello", so omitting a prim path in AddReference() targets the referenced layer’s default prim. NVIDIA’s reference documentation explains that a reference may omit the prim path when the referenced layer has a default prim, and the target prim will be resolved from that metadata. ( docs.nvidia.com )

Option A is incorrect because the reference is not authored under /hello; /hello exists only inside the referenced file. Option B is incomplete because it omits the reference metadata. This aligns with Composition → References, Default Prim, Layer Authoring, and Exported Layer Opinions .

usdchecker is important in data exchange because it validates whether exported USD or USDZ assets comply with rules that make them more likely to be interchangeable, renderable, and usable by downstream consumers. NVIDIA’s Learn OpenUSD Asset Validation lesson states that usdchecker validates a USD stage or USDZ package using defined rules to provide the best assurance that an asset will be properly interchangeable and renderable by Hydra. It is presented as part of the data exchange workflow for testing converter output and improving exporter quality.

Option A is correct because asset validation directly supports downstream consumption, which is the core goal of data exchange. A converter may write syntactically valid USD but still omit important metadata, invalid extents, default prims, unresolved dependencies, or other requirements that reduce interoperability. Option B is incorrect because usdchecker does not require binary .usdc output. Option C is incorrect because validation is not optimization. Option D is too narrow; URI and asset-path checks may be part of validation, but they are not the primary purpose. This aligns with Data Exchange → Asset Validation, usdchecker, Converter Testing, USDZ Compliance, and Downstream Interoperability .

The correct UsdLux light type is DistantLight . NVIDIA’s Learn OpenUSD lighting guide identifies UsdLux.DistantLight as a schema where light is emitted from a distant source along the negative Z axis, commonly known as a directional light. This makes it the most appropriate light type for representing sunlight in an outdoor scene.

Option C is correct because sunlight is effectively modeled as coming from an extremely distant source, so its rays are treated as broadly parallel across the scene. The OpenUSD DistantLight schema documentation explicitly describes it as typically used for distant, broad light sources like sunlight, where rays affect the entire scene and are roughly parallel. It also notes that the inputs:angle attribute controls angular diameter; the fallback value approximates the Sun’s apparent angular size from Earth.

Option A , RectLight, is better suited for rectangular area emitters such as panels or windows. Option B , DomeLight, is commonly used for environment or HDRI lighting. Option D , SphereLight, represents local emission from a spherical source, such as a bulb. This aligns with Visualization → UsdLux, Lighting Schemas, DistantLight, Directional Lighting, and Outdoor Illumination .

Setting instanceable = false on an instance root is a deinstancing operation. NVIDIA’s Learn OpenUSD instancing material explains that instanceable prims share composed data through implicit prototypes, and that scenegraph instances restrict sparse descendant overrides because instance proxies are not directly editable. In NVIDIA’s instance-refinement guidance, the three practical rules are: prototypes are not editable, instance proxies are not editable and local opinions are discarded, while the instanceable prim itself remains editable. ( docs.nvidia.com )

Therefore, when the targeted instance root is deinstanced, local opinions authored under that prim’s descendant paths can become effective because the subtree is no longer represented only through the shared prototype. USD must also recompose that hierarchy because changing instanceable changes whether the prim contributes to prototype sharing or composes its own local subtree. Option A is incorrect because disabling instanceable on one prim does not mutate metadata on other prims that previously shared the prototype. Option C describes the still-instanced behavior, not the deinstanced result. In the screenshot, the stem names /ParkingLot/Car_2 while options B and D name /ParkingLot/Car_1; the correct effects apply to the prim whose instanceable metadata is disabled. This aligns with Content Aggregation → Scenegraph Instancing, Instance Refinement, Deinstancing, and Recomposition .

The immutable stage-opening concerns are the path resolver context and the variant fallback configuration used for unselected variants . The resolver context is bound when the stage is created or opened and is used for future asset-path resolution on that stage, regardless of what other resolver context may be bound elsewhere. The OpenUSD UsdStage API documentation describes this binding behavior during stage creation, making option C correct.

Option D is also correct because global variant fallback preferences are defined as preferences “used in new UsdStages.” Once a stage has been composed with its fallback preferences, changing global fallback settings affects newly opened stages, not the already-opened stage’s established fallback behavior.

Option A is incorrect because payload loading is mutable: Load(), Unload(), LoadAndUnload(), and SetLoadRules() modify the stage’s payload working set after opening. Option B is incorrect because layers can be muted and unmuted on an existing stage through MuteLayer(), UnmuteLayer(), and MuteAndUnmuteLayers(). This aligns with Pipeline Development → Stage Opening, Asset Resolution, Variant Fallbacks, Load Rules, and Layer Muting .

A payload is most appropriate when the scene contains heavy assets that should not be fully composed and loaded until needed. NVIDIA’s Learn OpenUSD payload lesson describes payloads as similar to references, but with the added ability to load and unload them on demand. It gives the example of opening a large city scene with all payloads unloaded so the stage loads quickly, then selectively loading only the city block or asset of interest. When loaded, the payloaded layers and their composed dependencies are brought into the scene.

Option B is correct because payloads are designed for deferred loading, scalable working sets, reduced initial load time, and lower memory cost. Option A describes versioning or asset-resolution policy, not the primary distinction between payloads and references. Option C does not justify a payload; small inexpensive assets are often fine as references. Option D describes the opposite of payload behavior, because references are composed immediately, while payload traversal can be deferred. This aligns with Composition → References and Payloads, Load/Unload Behavior, Composition Arcs, Scalability, and LIVERPS Strength Ordering .

Usd.Property.GetPropertyStack() is the most appropriate choice because the debugging goal is to determine where a composed property value is coming from. In USD, a final property value can be affected by multiple authored opinions across sublayers, references, payloads, variants, inherits, specializes, and local overrides. NVIDIA’s Learn OpenUSD glossary defines a property stack as the ordered list of property specs that contribute values or metadata for a composed property, and notes that UsdProperty::GetPropertyStack() is specifically useful for debugging, not as the primary value-resolution mechanism. It also explains that USD resolves values by traversing composition information in strength order.

Option C is therefore correct because it exposes the contributing property specs and their layer origins, helping identify why a stronger opinion produced an unexpected result. Option A only traverses prims on the composed stage and does not explain value provenance. Option B reports aggregate stage statistics, not composition-source details. Option D exports the stage or layer contents but does not provide targeted diagnostic ordering for a property. This aligns with Debugging and Troubleshooting → Value Resolution, Property Stack, Composition Debugging, and Layer Opinion Inspection .

API schemas and typed schemas serve different modeling roles in OpenUSD. NVIDIA’s Learn OpenUSD schema guidance states that IsA, or typed, schemas define what a prim fundamentally is by assigning a typeName, while API schemas define what capabilities a prim has by adding optional properties or behaviors to an already-typed prim. API schemas do not assign a typeName; instead, they are list-edited through the apiSchemas metadata and queried with HasAPI. ( docs.nvidia.com )

Option A is correct because API schemas can be applied across different prim types, while a typed schema defines a specific prim type such as Mesh, Camera, or Xform. Option B is also correct because one prim has a single typed schema identity, but multiple API schemas can be applied to enrich that prim with additional behaviors. NVIDIA further notes that API schemas may be single-apply or multiple-apply, where multiple-apply schemas can be applied more than once to the same prim using different instance names. ( docs.nvidia.com )

Option C is incorrect because API schemas can define attributes and relationships. Option D is incorrect because Python binding policy is not the distinguishing factor between API and typed schemas. This aligns with Customizing USD → Schemas, IsA Schemas, API Schemas, typeName, apiSchemas Metadata .

The correct method is my_prim.GetAttribute("size") because the question asks for a specific attribute , not a relationship or generic property. NVIDIA’s Learn OpenUSD material defines properties as the data-bearing namespace objects on prims and distinguishes two property categories: attributes and relationships . Attributes hold typed values, while relationships point to other objects in the scene description. NVIDIA’s Omniverse developer reference also states that Usd.Prim.GetAttribute() returns a Usd.Attribute, after which Usd.Attribute.Get() is used to resolve the actual authored or composed value.

Option A is therefore the precise API call. Option B is incorrect because GetRelationship() retrieves relationship properties, not value-bearing attributes. Option C is less specific: GetProperty("size") can retrieve either an attribute or relationship as a generic UsdProperty, requiring additional type handling. Option D is incorrect because primvars are accessed through schema-specific APIs such as UsdGeom.PrimvarsAPI, not directly as a general Usd.Prim method. This aligns with Pipeline Development → USD Python API, Property Access, Attributes, Relationships, and Scenegraph Authoring .

An authored opinion may be valid and still not appear in the final composed stage because USD resolves competing opinions by strength ordering . NVIDIA’s Learn OpenUSD strength-ordering guide explains that, for each prim and property, the composition engine evaluates opinions from contributing layers and composition arcs using LIVRPS, giving precedence to stronger opinions when conflicts occur. It also states that stronger layers can override or add to data defined in weaker layers, enabling non-destructive edits and overrides.

Option D is correct because a weaker layer’s opinion can be hidden by a stronger opinion authored elsewhere in the layer stack or through a stronger composition context. This is a normal result of USD composition, not an error. Option A is incorrect because editability controls whether a layer can be modified, not whether its authored opinions can contribute during composition. Option B may indicate invalid or poorly modeled data, but schema choice alone is not the general reason an opinion loses. Option C can be a troubleshooting case for missing referenced content, but the question asks why opinions from a composed layer may not take effect. This aligns with Composition → Layer Stacks, Opinion Strength, LIVRPS, Value Resolution, and Non-Destructive Overrides .

The correct UsdShade prim type is NodeGraph. NVIDIA’s Learn OpenUSD material introduces UsdShade as the schema family used for creating and binding materials, where materials store renderer-facing shading definitions. The OpenUSD UsdShade specification defines UsdShadeNodeGraph as the mechanism for packaging shading networks into reusable units. A node graph can contain shader nodes and nested node graphs, expose public inputs as an interface, provide outputs, and be referenced into larger networks as a reusable building block.

Option C is correct because NodeGraph is explicitly designed for reusable and parameterized shading-network structure. It allows a pipeline to encapsulate repeated shading logic while exposing only selected parameters to consuming materials or larger graphs. Option A is incorrect because SubNetwork is not the correct UsdShade prim type in OpenUSD. Option B is incorrect because a custom schema could theoretically define new data models, but it is not the standard UsdShade construct for reusable shader-network packaging. This aligns with Visualization → UsdShade, Materials, Shading Networks, NodeGraph Interfaces, Parameterization, and Reusable Material Components .

The correct troubleshooting path is to verify the prim’s kind and whether UsdGeomModelAPI behavior is properly applied. OpenUSD’s UsdGeomModelAPI documentation states that draw modes provide alternate imaging behavior for USD subtrees with kind model . The attributes model:drawMode and model:applyDrawMode are resolved to decide whether traversal should stop at a model boundary and replace the subtree with proxy geometry. For bounds, the replacement is the model-space bounding box of the replaced prim. ( openusd.org )

Option A is correct because drawMode = "bounds" is not a generic visibility toggle for arbitrary prims. It is a model-level imaging mechanism. The prim must participate correctly in the model hierarchy, and the relevant UsdGeomModelAPI attributes must be authored or inherited in a way that causes draw mode application. The documentation also notes that component models are automatically treated as if model:applyDrawMode were true unless explicitly disabled. ( openusd.org )

Options B , C , and D are unrelated to model draw-mode activation. Physics collision APIs, volume schemas, and material binding APIs can affect simulation, volume representation, or shading, but they do not control whether model draw modes are applied. This aligns with Visualization → Model Draw Modes, UsdGeomModelAPI, Kinds, Bounds, and Imaging Substitution .

At timeCode 45, the referenced animation from layer1.usda interpolates between the samples at 30 and 60. Since the Y values are -5.0 at frame 30 and 5.0 at frame 60, the interpolated local translate at frame 45 is 0.0. Option A works because adding a parent transform on /World at frame 45 contributes an additional Y translation of -5.0; combined with the sphere’s interpolated local value of 0.0, the resulting placement is Y = -5.0.

Option C also works because a layer offset retimes animation across a reference. NVIDIA defines a layer offset as an adjustment to time values when composing layers through references, payloads, or sublayers, using offset and scale to retime animated data non-destructively. With offset = 15, the source sample at frame 30 is composed at frame 45, so the referenced sphere evaluates to Y = -5.0 at root time 45. Value resolution accounts for layer offsets and interpolation of time samples.

Option B only changes timeline metadata and does not retime or override the animation. Option D interpolates between -2.5 at frame 30 and -5.0 at frame 60, producing -3.75 at frame 45, not -5.0. This aligns with Composition → References, Layer Offsets, Time Samples, and Value Resolution .