Evertz Positions DreamCatcher System for Video Production Powered by Real-Time, AI-Built Metadata
Partnership with ShotTracker, Ease Live uses data to automate storytelling
In an era when analytics are driving on-field decisions of the biggest professional sports franchises, it’s only a matter of time before real-time data driven by artificial intelligence fuels live and on-demand content creation by sports broadcasters.
Evertz believes that time is here with the latest developments behind its IP-based production suite, DreamCatcher.
Through a partnership with sensor-based system ShotTracker and the use of technology obtained in Evertz’s acquisition last year of interactive graphics platform Ease Live, Evertz is targeting the collegiate-athletics market with a platform it describes as “data-driven video production.”
Essentially, the system uses tracking data, a locked-off 4K camera (or other autonomous cameras), and timecode synchronization with all the cameras in a broadcaster’s arsenal to tie everything together to build replays and curated highlight packages built on real-time metadata.
The system turns data into a tool in the storytelling arsenal. Want to go into a game tracking one player’s rebounds? Set the metadata, and a highlight playlist is built using angles from the locked-off camera as well as all the various traditional broadcast feeds that a timecode synched into the system. Did a surprising player off the bench have a big day? Add in metadata fields during or after the game, and DreamCatcher will mine the data already collected to spit out the exact content that you want.
“All of that metadata, whether it’s real-time or non–real-time, adds to the backend, where you can do live production or automated live production,” says Nima Malekmanesh, product marketing manager/senior engineer, DreamCatcher, Evertz. “You can add to and augment your live production; your traditional replay operators won’t have to clip content, name it, tag it, export it. Referees would automatically see plays [in Video Assistant Refereeing] from multiple angles. It can all be automated.”
According to Evertz, the platform is valuable to live-content creators in three areas: highlight creation (where producers can mine the data for stories), clip creation (for operators who need to focus in more than one area when cutting replays for a live production), and automated logging and exporting to a broadcaster’s asset-management software.
In the live environment, the ability to deploy the 4K camera to zoom for automated camera follows is also a valuable add. Using digital zoom, DreamCatcher can isolate and follow a specific player or players on command, providing replays to roll based on predetermined storylines from a non-operated camera. For on-demand content, automated clipping will also immediately cut and appropriately tag and log a play in real time. All while synchronizing the locked-off camera source with the other broadcast cameras to build more-dynamic replays, highlight packages, and even melts.
“Because we have real-time X,Y,Z data coming in through ShotTracker, with the follow of the ball, I can select multiple players,” says Malekmanesh. “The image will then zoom in and out, [keeping] those players in the [shot] at all times. It’s very powerful [from the perspective of] a coach or a fan or a broadcaster. Maybe there’s a matchup between a receiver and a cornerback that you want to follow. You can tailor and have as many of these isos as you want.”
Through Ease Live, the system can also make it possible to put the controls in the hands of the viewer. Through streaming distribution models, broadcasters could offer fans the ability to custom-build their own highlight packages or real-time camera follows through live interactive graphics built directly into the live stream’s video.
According to Evertz, more than 50 colleges, many in the SEC and ACC, are already DreamCatcher customers who have access to this technology. The full platform was given its first real end-to-end test at the Big 12 Men’s Basketball Tournament in March. Currently, the platform is most naturally deployed for basketball, but the company says football and even volleyball aren’t far behind as AI continues to grow smarter in those sports.
The technology is rapidly maturing, according to Evertz engineers, who note that, currently, the biggest challenges lie in installation of workflows on the end-user side.
“Getting production teams to adopt a new workflow [is a challenge],” says Malekmanesh. “That’s where the challenges lie; the technology is actually very resilient. Getting the schools to wear the sensors, [too].”
The technology offers highly interesting prospects for a future where analytics won’t just be forcing baseball teams into infield shifts but will be dictating the content-creation methods for sports-content rightsholders big and small.