From standards to public software: Nokia advances AI-enhanced video for all

The next leap in video quality will not come from compression alone. It will come from AI that enhances what viewers see after a video is decoded. Nokia’s public Neural Network Post Filtering (NNPF) software release is a practical demonstration of how AI-based post-filtering can deliver better, more efficient and scalable streaming experiences.

Most people never think about codecs, metadata or AI inference libraries. They notice something simpler: whether the video they watch looks crisp, natural and immersive. When fine textures disappear, gradients break apart or dark scenes lose detail, the user experience feels less convincing. NNPF helps close that gap by applying AI at the final stage of the video pipeline, improving the image viewers see while staying aligned with standards-based innovation the industry depends on.

This release builds on our earlier work on neural network post-filtering within the wider VVC and VSEI standards ecosystems. In simple terms, NNPF uses AI after decoding a video, to restore detail, sharpen textures and deliver improved perceptual quality without needing increased bitrate. For streaming platforms and device makers, that means a more efficient path to better picture quality. For viewers, it means video that looks richer and more lifelike on the devices they already use.  

Why this release is important now

Timing matters. VVC is emerging as a foundation for next-generation video services, delivering higher quality at lower bitrates across broadcasting and streaming use cases. At the same time, consumer devices are gaining dedicated AI acceleration, from laptops and phones to TVs and set-top devices. Our public NNPF software release sits squarely at that intersection: a more efficient codec on one side, and practical AI execution on everyday hardware on the other.

That makes the release more than a code drop. It is a concrete demonstration of how standards, software and hardware can advance together. The software described here combines a VVC decoder path with support for NNPF SEI messages, neural network weight handling and real-time inference through OpenVINO. In other words, it shows how AI-enhanced video can be delivered in a standards-aligned, interoperable way rather than as a closed one-off demo.
 

 

Figure 1.  From VVC bitstream to enhanced picture.  Nokia’s public NNPF software release demonstrates a practical, standards-aligned workflow for AI-enhanced video, combining VVC decoding, synchronized metadata and neural network inference in one interoperable toolchain. 

Why NNPF matters for consumers

For viewers, the benefit is straightforward. Better compression and better post-processing mean cleaner, more immersive video even when bandwidth is limited. Fine textures hold up better. Edges look more natural. Skies, shadows and subtle gradients retain more of the detail that often gets lost in delivery. And because NNPF works after decoding, these gains can be achieved without redesigning the entire distribution chain.

That matters whether someone is watching premium drama on a big screen, sports on a laptop or short-form video on a phone. It also matters where networks are constrained. If the same bitrate can deliver a better visual experience, streaming becomes more resilient, more efficient and more inclusive. As we highlighted in our earlier NNPF work, AI can help restore the details compression removed, delivering richer colors, smoother gradients and sharper textures on consumer devices already in the market. 

Figure 2. Better video quality without more bitrate. Sample results from Nokia’s NNPF workflow show measurable quality gains, illustrating how AI-based post-filtering can improve perceived video quality without increasing delivery bitrate.

What NNPF means for the video industry

For the media and entertainment ecosystem, the implications are broader. Public software lowers the barrier to experimentation. Researchers, developers, encoder vendors and platform teams can now explore a concrete implementation path for AI-based post-filtering with VVC. The NNPF software is available under the BSD 3-Clause Clear License to support evaluation and experimentation, while use of any underlying patented Nokia inventions remains subject to Nokia’s applicable licensing programs. This release demonstrates the practical benefits of AI-based post-filtering and help create a clearer path toward future adoption. That accelerates learning, interoperability testing  and productization. Just as important, it helps move the conversation from what AI-enhanced video could do to what it can already do in practice.

This is especially relevant as broadcasters and streamers look for new ways to stand out. The Media Coding Industry Forum (MCIF) has highlighted how VVC supports next-generation services for broadcasting and streaming, from better efficiency to more immersive experiences. Our NNPF software release adds another important layer to that story: not only coding efficiency, but AI-assisted visual enhancement delivered through a standards-based workflow.  
 

 

Figure 3. AI video enhancement on consumer-class hardware. Real-time NNPF inference on modern consumer hardware highlights how dedicated AI acceleration is making standards-based video enhancement practical at scale.

From research leadership to real-world momentum

Nokia has helped shape video technology for decades, from contributions to successive generations of video codecs to more recent leadership in the AI era of multimedia. Our work on VSEI is especially important because it extends established H.26x video standards with synchronized metadata and enhancement capabilities, allowing video systems to evolve without breaking compatibility. That is exactly the kind of bridge the industry needs as AI becomes part of mainstream media pipelines.  

This public software release strengthens that position. It shows that Nokia is not only helping define the standards path but also translating those ideas into working software that others can inspect, test and build on. In a field as complex as AI-enhanced video, that combination of standards leadership and implementation leadership can make a real difference to adoption.

Explore the public NNPF software release and test it for yourself on GitHub: https://github.com/nokia/fraunhoferhhi-vvdec/blob/master/nnpf/README.md