Sony's new OCELLUS enables stable marker-free tracking with multiple sensors and seamless camera integration for cinematic and broadcast applications.
Sony is launching its first camera tracking system, OCELLUS, which has been designed for augmented reality and virtual productions in both broadcast and cinema by providing marker-free camera tracking through multiple sensors.
The system comprises a sensor unit, a processing box, and three lens encoders, and is camera agnostic and can be used with both cinema and broadcast cameras from Sony and elsewhere (though Sony cameras look to increase some functionality).
With the five image sensors and Sony's Visual SLAM (Simultaneous Localization and Mapping) technology, the system creates a reference map, enabling stable marker-free tracking both indoors and outdoors.
When using Sony cameras, metadata regarding focus, iris and zoom values from the camera and lens can be obtained via the camera's SDI output and transmitted in real-time to external devices via Ethernet cable. Currently this supports Cooke /i lenses, B4 and E-mount lenses. If the lens does not support metadata acquisition through the camera, lens encoders can be affixed to the camera to obtain this metadata.
The system also supports recording tracking data, camera/lens metadata, timecode and file name, which can be used for the post-production workflow.
Here's how it looks when mounted with an explanation of each of the three components.
1. Compact and lightweight sensor unit with five image sensors:
- Four image sensors out of five on the sensor unit are selected to use, providing stable marker-free tracking and high occlusion resistance, critical for operations
- If at least one image sensor in use captures valid feature points, tracking data can be extracted
- IR LEDs on both sides of each image sensor help tracking in low-light environments
- Visible Light Cut Unit included for stable tracking in environments with frequent lighting changes
- Sensor unit dimensions: approx. 86 mm × 60 mm × 43 mm (W × H × D) (3.39" × 2.36" × 1.69"), weight: approx. 250 g
- Easy installation and position adjustment using the NATO rail mounting parts (included)
- Connection to the processing box via a single USB Type-C cable with a lock mechanism, powered by the processing box via same USB Type-C cable
2. Processing Box:
- Real-time transmission of tracking data and camera2/lens3 metadata to CG rendering software like Unreal Engine via Ethernet cable[ in free-d format
- Equipped with Genlock input, Timecode input, SDI input/output terminals, and lens encoder connection terminals
- Supports recording tracking data and camera/lens metadata as FBX files on SDXC memory cards (UHS-II/UHS-I) synchronized with video files of main camera
- OLED display for checking IP address, tracking information, lens data, and more
3. Lens Encoder:
- Detects precise rotation angles and positions of lens focus, zoom, and iris values
- Transmits detected data to the processing box via LEMO 7-pin cable
- Enables metadata acquisition for lenses and cameras not supporting lens data embedding on SDI output
- Includes five different types of gears for various lenses
Pricing and availability
No word yet on pricing, but the projected release date for the system is November 2025 and it will be on show at NAB.