In addition to intelligent 16×9 to 9×16 cropping of live video, Inference automatically detects and extracts clips from live content for real-time highlight delivery.
AWS announced AWS Elemental Inference, a fully managed AI service that automatically transforms live and on-demand video broadcasts for mobile viewing audiences. The AI-powered cropping tool intelligently converts 16×9 landscape broadcasts into a vertical 9×16 format optimized for social and mobile platforms. The service automatically reformats content for mobile viewing while maintaining broadcast quality by tracking subject matter and keeping key actions visible.
AWS Elemental Inference allows broadcasters and streamers to reach audiences on social and mobile platforms like TikTok, Instagram Reels, and YouTube Shorts without manual post-production work or AI expertise.
Audiences today consume different content than they did a few years ago. However, most live broadcasts are produced in landscape format for traditional viewing. Converting these broadcasts into vertical formats for mobile platforms typically requires time-consuming manual editing. As a result, broadcasters miss out on viral moments and lose viewers to mobile-first destinations.
The service uses agent AI applications to analyze videos in real-time and automatically apply the right optimizations at the right time. Vertical video crop detection and clip generation occur independently, performing a multi-step transformation that requires no human intervention to extract values.
AWS Elemental Inference analyzes videos and automatically applies AI features without the need for human intervention. While creators focus on producing high-quality videos, AWS Elemental Inference autonomously optimizes content and provides a personalized content experience for viewers.
AWS Elemental Inference applies AI capabilities in parallel with live video, delivering latencies of 6 to 10 seconds, compared to the minutes required with traditional post-processing approaches. This “process once, optimize everywhere” approach allows multiple AI functions to run simultaneously on the same video stream, eliminating the need to reprocess content for each function.
Inference also automatically detects and extracts clips from live content highlights for real-time distribution. For live broadcasts, this means identifying game-winning plays, touchdowns, and emotional peaks with exact start and end points, cutting manual editing efforts from hours to minutes.
The service seamlessly integrates with AWS Elemental MediaLive, allowing users to enable AI capabilities without changing their existing video architecture. AWS Elemental Inference uses a fully managed foundational model (FM) that automatically updates and optimizes, eliminating the need for a dedicated AI team or expertise.
AWS Elemental Inference is currently available in four AWS Regions: US East (N. Virginia), US West (Oregon), Europe (Ireland), and Asia Pacific (Mumbai). AWS Elemental Inference can be enabled through the AWS Elemental MediaLive console or integrated into your workflow using the API. Pay-as-you-go pricing means you only pay for the features you use and the videos you process, with no upfront costs or contracts. This means you can scale during peak event periods and optimize costs during slow periods. Keep an eye on this space as more features will be introduced throughout this year, including deeper integration with core Elemental services and the ability to help customers monetize their video content.
Stay tuned to SVG later this week to learn more about the AWS Elemental Inference platform. In the meantime, please visit the AWS Elemental Inference product page and the AWS Elemental Inference documentation for more information.
