Skip to main content

C++ Visualization Tools

This page covers the C++ visualization tools available in Project Aria Tools. These tools provide native performance visualization of Aria data using the Pangolin framework.

Prerequisites​

To use the C++ visualization tools, you need to build them from source using CMake. These tools are not included in the standard Python package installation.

Required build step: Follow the "build using CMake from source code" instructions in the Advanced Installation guide to compile the C++ visualization tools.

The C++ tools require the Pangolin library for visualization. If Pangolin is not found during compilation, the visualization tools will not be built.

aria_viewer​

The aria_viewer is a C++ binary that provides real-time visualization of Aria VRS (Video Recording and Sensor) data using the Pangolin framework. It offers native performance for visualizing multi-modal sensor data from Aria devices.

Docusaurus themed imageDocusaurus themed image

Basic Usage​

First, navigate to the built directory:

cd ${SRC}/build/tools/visualization/

Then run the viewer command:

aria_viewer --vrs path/to/your/file.vrs

Command Line Options​

ParameterTypeRequiredDescription
--vrsstringYesPath to the VRS file you want to visualize

What You'll See​

The C++ viewer provides an interactive visualization environment with multiple components:

  • RGB & SLAM Camera Streams: Real-time display of RGB and SLAM camera images with overlaid eye gaze and hand tracking results
  • 1D Sensor Data: Time series plots of:
    • IMU signals (accelerometer and gyroscope data)
    • Audio signals (microphone waveforms)
    • Barometer signals (pressure and temperature readings)
  • 3D Visualization: Interactive 3D scene showing:
    • VIO trajectory (Visual-Inertial Odometry path)
    • Eye gaze vectors in 3D space
    • Hand tracking results and poses
    • Device pose and orientation

Interactive Controls​

The viewer includes real-time controls for:

  • Play/Pause: Start and stop data playback
  • Speed Control: Adjust playback speed using a slider
  • Timestamp Navigation: Jump to specific timestamps by dragging the timestamp slider
  • Stream Toggles: Enable/disable individual data streams for focused analysis
  • 3D Camera Controls: Navigate and explore the 3D visualization

Known Issues and Workarounds​

High-frequency sequence playback issue: There is a known issue when playing high-frequency Aria-Gen2 data at real speed, where camera views may update asynchronously. This is due to CPU-based image decoding speed limitations.

Workarounds:

  1. Use Python viewer: Fall back to the aria_rerun_viewer for smooth high-frequency playback
  2. Reduce playback speed: Use the speed control slider to play at slower speeds
  3. Random access: Dragging the timestamp slider for random access always works correctly

The development team is actively working on a solution for this issue.