An encoder that understands what it streams.

Challenge

Broadcast television is moving on two axes at once: from HD to 4K, and from 4G to 5G networks. Video flows through encoders into H.264/H.265 streams and on to local or cloud storage — and with it comes a mountain of data to store and convert, plus the manual labour of tagging and labelling hours of content for later editing and playback. Meanwhile, AI edge computing has matured to the point where video-to-text conversion is practical in real time. Encoding and AI are no longer separate domains; they are merging into one.

Solution

YUAN's next-generation 4K60 broadcast encoder integrates all three requirements — top-tier compression, real-time streaming, and AI recognition — in one system. The NVIDIA Jetson Orin NX pairs with YUAN's SC710 capture cards for genuine 4K60P broadcast-grade capture: 4:4:4 chroma subsampling, 10-bit depth, and VANC capture, with NVIDIA GPUDirect RDMA cutting overall capture latency to 15 microseconds. Orin NX adds AV1 compression — quality comparable to H.265 at lower bitrates — trimming network load and slashing storage and transmission costs. YUAN's NexVDO SDK carries the industry's transmission standards, including RTSP, RTMP, TS, NDI, SRT, and Dante AV, alongside its own SkyLink X ultra-low-latency protocol.

The AI is what changes the category. With Orin NX modules offering 70 or 100 TOPS, the encoder runs AI and generative AI algorithms during the encode itself, extracting information from the video and writing it into the stream's metadata. Post-production teams search for the segment they need instead of scrubbing for it, and important timestamps get marked as they happen — a redefinition of what the next-generation encoder is for.

Key benefits

Efficient AV1 compression lets broadcasters use bandwidth judiciously without sacrificing visual quality, with considerable storage savings to match. The NexVDO SDK's protocol support — NDI, SRT, Dante AV, and CBR streaming for stable network rates — makes integration with existing broadcast equipment straightforward. And through SIE, Stream Information Extraction, audio encoding and AI-generated information compress in synchronisation, dramatically lowering the threshold for developers building generative AI into the broadcast chain.