LAS VEGAS - April 6, 2025 - PRLog -- At this year's NAB Show, Nokia, a member of the Media Coding Industry Forum, will demonstrate the latest advancements in video compression and enhancement leveraging artificial intelligence to deliver higher quality video at lower bitrates—without additional transmission cost or device complexity.

According to Sebastian Schwarz, Head of Video Coding Systems at Nokia Technologies, the demonstration will highlight the integration of Versatile Video Coding (VVC) with Versatile Supplemental Enhancement Information (VSEI) and real-time neural network post-filters—an innovative combination that promises to significantly improve end-user video experiences for streaming service providers, broadcasters, and device manufacturers.

What's on Display?

Nokia's demonstration shows how neural network filters—executed in real-time on commercial off-the-shelf hardware—can dynamically enhance video quality after decoding. These AI-powered filters, designed to restore visual detail and improve image fidelity, are guided by real-time metadata updates embedded in the video bitstream via the VSEI standard.

For this showcase, Nokia is using a standard Intel Lunar Lake laptop equipped with a built-in Neural Processing Unit (NPU) and hardware VVC decoding capabilities. This setup enables real-time application of neural network post-filters directly on the playback device—without requiring any specialized or proprietary hardware.

A Real-Time Leap in Efficiency

The practical impact is striking: Nokia demonstrates how content can be transmitted at a slightly lower quality—saving approximately 5% in bitrate—while still delivering perceptually equivalent visual quality to viewers using these AI filters. This lightweight enhancement method adds no overhead to systems that do not support it, and gracefully scales quality for those that do.

Why It Matters

This technology empowers streaming platforms and content providers to:

• Improve visual quality dynamically across different content types and scene transitions.
• Reduce bandwidth consumption without compromising user experience.
• Add next-gen enhancements like super-resolution (e.g., upscaling from HD to 4K) and temporal resolution boosts (e.g., from 30fps to 60fps), all through flexible metadata-driven neural network updates.

• Streaming service providers looking to optimize delivery quality and efficiency.
• Broadcasters striving to stay competitive in an IP-based distribution future.
• Device manufacturers seeking to showcase enhanced capabilities in next-gen TVs, smartphones, or media players.
• Content developers and distributors interested in adding value to their media without reengineering workflows.