NVIDIA ovrtx

Omniverse RTX is the technology that provides real-time, physically accurate sensor simulation and rendering for Physical AI, targeting robotics learning, synthetic data generation, and industrial and design workflows. ovrtx is the lightweight C and Python SDK that exposes Omniverse RTX—you use it to integrate that sensor simulation and visualization into your own applications.

In this documentation you will find getting started guides for Python and C, API references, and example projects.

• Getting Started in Python

• Getting Started in C

Note

ovrtx is currently pre-release software.

Features

• Physically accurate simulation of cameras, lidar, radar, and other sensors.

• Scalable simulation performance from reinforcement learning in-the-loop with tens of thousands of frames per second, through real-time, photorealistic, interactive viewport and navigation, to offline predictive rendering.

• OpenUSD scene description allowing interchange with a vast ecosystem of content creation, CAD and simulation tools.

• Easy integration with Python simulation and learning ecosystem.

Support

https://forums.developer.nvidia.com/c/omniverse/300

License

The software and materials are governed by the NVIDIA Software License Agreement and the Product Specific Terms for NVIDIA AI Products.

Getting Started

• Getting Started in Python
  • Minimal Example
  • Next Steps
• Getting Started in C
  • Installation
  • Minimal Example
  • Next Steps

API Usage

• Application Flow
  • Minimal Lifecycle
  • Python vs. C
  • Where to Go Next
• Renderer Configuration
  • Creating a Renderer
  • Configuration Entries
  • Runtime Package Layout
  • Multi-Renderer Processes
  • Cleanup
• Async Operations, Status, and Errors
  • Python Operations
  • C Waits
  • Status Queries
  • Error Handling
  • C Log Callback

Sensors

• Sensor Configuration
• Sensor Outputs
• Cameras
  • Render Modes
    • Real-Time Path-Tracing
    • Path Tracing
    • Minimal
  • Outputs
• Non-Visual Sensors
  • Lidar Sensors
  • Radar Sensors
  • Reading Sensor PointClouds
  • Non-Visual Materials

Scene

• Loading USD
  • Opening a Root Layer
  • Inline Composition
  • References
  • Time-Sampled USD
  • Resetting the Stage
  • Authored Attribute Population
• Stage Queries
  • Python Queries
  • C Queries
  • Async Queries
  • Path Dictionary
  • Troubleshooting
• Attribute Reads and Writes
  • Tensor Layout
  • Reading Attributes
  • Writing Attributes
  • Data Access
  • Type Notes
  • C Convenience Helpers
  • Troubleshooting
• Writing Transforms
  • 4x4 Matrix Writes
  • C Transform Helpers
  • Repeated Updates
  • Notes
• Attribute Bindings
  • Create and Write
  • Async Creation and Writes
  • Array Attributes
  • Mapping Through a Binding
  • Lifetime Rules
• Attribute Mapping
  • CPU Mapping
  • CUDA Mapping
  • Explicit Unmap
  • Limits and Lifetime
• Cloning Prims
  • Clone a Subtree
  • Async Cloning
  • Notes
• Material Binding
  • Binding a Material
• Semantic Labels
  • Author Labels in USDA
  • Runtime Overrides
  • Interpreting Segmentation Output
  • Notes
• Picking and Selection
  • Concepts
  • Interactive Viewport Workflow
  • Pick Rectangles
  • Pick Results
  • Resolving Picked Prim Names
  • Selection Outlines
  • Selection Styling
  • Pickable Prims
  • Reference

Examples

• Examples
  • Example Projects

Python API

• Python API Reference
  • ovrtx

C API

• C API Practical Patterns
  • Version and Configuration
  • String Handling
  • Error Lifetimes
  • Status Queries
  • Logging Callback
  • Resource Cleanup
• C API Reference
  • Overview
  • Functions
  • Types

Links

• GitHub Repository

Indices and Tables

• Index

• Search Page