- NHL uses AWS Elemental Link UHD encoders
- Puck-tracking system captures 1m data points a game
- AI and machine learning could unlock new possibilities in the future
The National Hockey League (NHL) is using cloud technology to capture and pair ultra-high definition (UHD) video with data gathered from its Puck and Player Tracking (PPT) system to create new digital experiences for fans and tools for players, coaches, and scouts.
The league appointed Amazon Web Services (AWS) as its official cloud, artificial intelligence (AI) and machine learning (ML) provider earlier this year, and is the first major sports organisation to deploy AWS Elemental Link UHD cloud encoders in its video production process.
These Link devices capture, process, and compress live video feeds in the cloud and eliminate the need for expensive and complex dedicated broadcast infrastructure. The NHL says traditional hardware was not a cost-effective solution when it has 190 different video channels and more than 1,400 games a season.
The league deployed UHD cameras and six Link devices at all 32 NHL arenas during the offseason and the setup is already unlocking a raft of use cases.
“Getting high quality UHD video from hundreds of cameras across dozens of disparate arenas via traditional hardware is neither practical, nor cost-efficient,” said Grant Nodine, head of technology at the NHL. “The plug and play nature of Link UHD, its affordable price point, and the way it allows us to use APIs to remotely control the devices without manual intervention or having to write new code offer a better alternative and have been game changing in that respect.”
The NHL can now produce UHD feeds for broadcast, its own video platform, and the in-venue replay system, while the speed of the process makes it feasible to combine high quality video with an on-screen graphic displaying real-time data collected from PPT, such as player speed, which is also hosted on AWS.
“We’ve spent a significant amount of time and capital into the puck-tracking system which generates one million datapoints every single game,” said NHL executive vice president of business development innovation Dave Lehanski. “This data is allowing us to tell a tremendous number of new stories, create new content, and offer a new perspective on the game. But we know to transform the fan experience with the data we need to transform the video experience. Not just for fans but our coaches, our officials, and media partners.
“Low latency is required for everything we do – we need to process this data and get it back in real time. We have so many audiences we’re trying to serve during a live game and our greatest asset, the speed of the sport, is also a challenge from a presentation perspective. Most stoppages in an NHL game are only a few seconds.
“Players will literally come off the ice and pick up an iPad to review video footage. There’s no other way [than the cloud] to get that video with the metadata [as rapidly]. We’re [also] spending a lot of time with our betting partners to build live in-play betting solutions where fans can make decisions using real-time data.”
It’s not just about live streams, either. The ultimate goal is to create a single video, data and application repository that makes it as simple as possible to share footage that can be used to improve pre and post-match production and to create on-demand content. Longer term, computer vision will make it possible to capture data from a range of live and archive video.
“It’s the installation of these new cameras and new encoders that are allowing this to happen,” Lehanski added.