Shulou(Shulou.com)11/24 Report--

The 25th Hi-Tech Fair, with the theme of "stimulating the vitality of innovation and improving the quality of development", kicked off in Shenzhen from November 15 to 19, when structural science and technology was unveiled as a "national specialty special new giant" enterprise. and released at the scene to build a new generation of "intelligent audio and video interaction engine", innovative audio and video real-time interactive new experience.

That is, the new generation of "intelligent audio and video interactive engine" has been completely upgraded from the underlying audio and video AI engine, audio and video cloud real-time transmission, audio and video ecological capabilities, and real-time interactive Homo sapiens. Qi Weibin, head of real-time interactive products, made a detailed interpretation at the high-tech fair.

Elegant demeanor of the on-site press conference

One-click access to a comprehensive solution to improve picture quality, providing the best picture quality according to the user's network status, mobile phone performance, scene and budget, and obtaining ultra-high definition picture quality. A self-developed video AI engine is constructed, which supports mobile super division, dark light enhancement, background segmentation, color enhancement, video noise reduction and other capabilities, optimizes coding and decoding, greatly improves coding efficiency, saves 30% or more bit rate, and supports ROI, layered coding, large and small streams and other capabilities.

The constructed audio AI engine is committed to providing users with a more vivid, interesting and immersive audio interactive experience. Self-developed 3A algorithm can obtain pure human voice after echo suppression under very low performance loss; self-developed scene AI noise reduction, intelligent adjustment of AI noise reduction strategy to provide the best noise reduction and sound quality effect, and in the music scene, can retain more music details while suppressing other noise. In addition, there are AI sound changes, reverberation, spatial audio and other effects to provide immersive sound quality experience.

In addition, the release of the new generation of "Intelligent Audio and Video Interactive engine" also makes comprehensive optimization for audio and video cloud real-time transmission. Based on massive and orderly network MSDN global coverage, it adopts the best capture rendering, codec strategy, to obtain the lowest delay, sound quality and picture quality, bringing optimal performance consumption, and enabling excellent performance of CPU utilization, memory usage, heating and power consumption of instant audio and video service. At present, the architecture can achieve that the first frame delay is lower than 400ms, the video stutter rate is less than 2%, and it is extremely resistant to weak networks. The end-to-end delay is as low as 70ms, the maximum packet loss is 80%, and the lowest bandwidth speed limit can reach 30kbps.

In order to provide customers with the best and most comprehensive services, the structure is also constantly building ecological capacity to allow customers to quickly access. Content risk control ensures business security and compliance, and works with intelligent voice industry ecological service providers to provide real-time voice transcription and translation services, so that users can communicate smoothly without fear of language barriers. Combined with partners' rich edge computing nodes and flexible GPU virtualization technology, the application client is cloud-based, and more cloud applications such as cloud rendering and cloud games are realized.

That is, digital Homo sapiens, as a self-developed AI digital human platform launched this year, can restore real human expression and expression, and can drive digital people to communicate and interact like real people through text or voice. The newly released "Real-time Interactive Digital Homo sapiens", based on the professional knowledge base of industry customers, provides real-time interactive solutions such as interactive screen, voice and text question and answer, real-time hosting and so on through large model and NLP, real-time push and pull stream and digital human image presentation.

That is, the release of a new generation of intelligent audio and video interactive engine will provide users with a more reliable and high-quality audio and video communication experience, and will continue to enable pan-entertainment, online education, remote conferencing, digital live broadcasting and other fields. In the future, it will continue to devote itself to the research and development and innovation of audio and video technology, and constantly improve product performance and user experience.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.