Volumetric Capture Application — Complete Rebuild (Dual Setup)

This release marks a complete rebuild of the volumetric capture application from the ground up. It maintains all core functionalities of the original program:

• Multi-camera streaming for live point cloud visualization
• Robust calibration workflow
• Real-time reconstruction and streaming

These foundational features are fully preserved. What sets this version apart is the new plugin-based architecture, making the application far more expandable and adaptable. The modular system enables easier integration of new features and extensions, ensuring long-term flexibility.

At the end-to-end workflow level, the Unity scene now supports:

• Volume rendering from two independent Volcap application instances at once
• Simultaneous audio streaming from both sources
• Seamless visualization and audio experience in a single Unity environment

This upgrade delivers the familiar power of the previous system, reimagined with expandability and dual-source capability for advanced volumetric capture workflows.

Documentation

This new release also comes with a new, organized documentation page at https://vcl3d.github.io/VolumetricCapture.
The goal was to consolidate the previous wiki and the great feedback and comments that were received so far.
Please make the effort to go through it when setting everything up and even contribute, or make suggestions through new issues, where necessary.

Major Updates

• Mixed device mode

  • Both Intel RealSense 2.0 D415 and Microsoft Kinect 4 Azure are supported in this release.
  • Connection, calibration and software synchronization are now possible with mixed devices. (Hardware synchronization is only support for the same type of devices unfortunately.)
  • Decoupled configuration for each type of device.

• Low-bandwidth streaming

  Connecting mulitple devices is usually hindered by the streaming bandwidth.
  This was the case especially for K4A whose MJPEG mode did not allow for configuring the compression parameters.
  With this new mode we drop the color resolution, while also enabling compression quality configuration to manage the overall bandwidth consumption and allow for real-time streaming when deploying a higher number of sensors.

• Updated Volumetric Sensor Alignment

  • Can now calibrate >= 3 sensors
  • Improved alignment quality
  • Supports both K4A and D415 (separately and mixed) in various depth resolutions

• Hardware synchronization support for K4A

  • Can now select a master for K4A as well
  • Automatic cascaded offset to avoid modulation frequency noise

Other fixes

• Resource cleanup errors upon closing
• First connection crashing the eyes
• Device repository documentation, multiple copies and mixed device issues
• Dumping of colored calibrated point clouds
• First time run GUI placement for volcap
• Can now run an eye on the same workstation running volcap
• K4A shader issue
• Depth units are now fixed for the two supported devices (100um for D415 and 1mm for K4A)

A new showcase section is now also available in the documentation where your deployments and works with the system can be hosted !

No major changes since release 4.0.1,
for more information, please refer to the update history

Features:

• Kinect Azure support.
• Updated volumetric calibration (see here) that supports more sensors at with greater placement variety.

Unfortunately, this early alpha release breaks backward compatibility with the previous version.
For Intel Realsense sensors support please use the previous release (3.0).

For more info visit our wiki under "Release 4.0"

• Volumetric Capture update
• Remote Eye update
• Multiview Player and Dumper
• Synchronization Validation Tool
• Device Repository