Big news! Kuaishou releases live streaming multi-bitrate adaptive industry standard

Recently, Kuaishou officially released the streaming-based live multi-bitrate adaptive standard LAS (Live Adaptive Streaming), which is used to provide a low-latency, smooth, and fluent live multi-bitrate experience. It is reported that Kuaishou also open-sourced the end-to-end solution of LAS, including the server, client, and industry-leading multi-bitrate adaptive algorithm, to help the industry achieve zero-threshold access and use of LAS.

The clarity of live broadcast is crucial to the user experience. By improving the bit rate and resolution of the video, the video clarity can be significantly improved. Kuaishou has a large user base and is widely distributed. The network differences between users are large. A single video quality (bit rate, resolution) or a fixed gear delivery strategy is difficult to adapt to different network requirements.

Kuaishou's technical team has developed a streaming-based live broadcast multi-bitrate technology that adaptively selects the best video quality for users based on their real-time network changes without adding extra live broadcast delay, providing users with a clearer and smoother viewing experience. This time, Kuaishou released the streaming-based live broadcast multi-bitrate adaptive standard LAS.

LAS has been in existence for nearly two years. After multiple iterations of optimization and polishing by Kuaishou engineers, it has gradually matured and has been fully deployed in Kuaishou's live broadcast business, significantly solving the pain points of live broadcast business such as video freezes, blurred image quality, and large live broadcast delays. Its stability and reliability have also successfully withstood the live broadcast test of large-scale events such as the National Day Celebration and the Spring Festival Gala last year. During this year's epidemic, the demand for live broadcast business has increased significantly. Kuaishou has benefited from the support of LAS technology in a variety of complex live broadcast business scenarios such as live e-commerce, online concerts, and online open classes.

During the epidemic, the National Grand Theater's "Spring Online" online series of concerts were live-broadcasted on Kuaishou, and the Kuaishou technical team provided a full technical solution from on-site shooting to transmission. The top-level viewing experience of "live broadcast + interaction + Dolby Theater Sound Effect" is matched with the playback clarity of "Blu-ray 8M"; the application of LAS adaptively selects the best video quality for users based on the user's real-time network changes without adding additional live broadcast delay, providing users with a clearer and smoother viewing experience. With the advantages of Kuaishou's live broadcast technology, classic art is promoted to the public.

It is understood that currently, domestic mainstream CDN vendors such as Alibaba Cloud, Baidu Smart Cloud, Kingsoft Cloud, Tencent Cloud, and Wangsu Technology all support the LAS standard and can directly provide LAS-based services. At the same time, the industry's well-known open source streaming media server SRS has also been upgraded to support the LAS standard. Based on SRS 4.0 and higher versions, enterprise customers can independently build LAS servers to meet personalized needs.

Dr. Zhou Chao, head of Kuaishou's transmission algorithm team, said that the purpose of releasing the LAS standard and opening it to the public is to open up Kuaishou's technology and experience accumulated over the years in the live broadcast field to the industry, help live video companies empower their own businesses through LAS, gain revenue faster, and help the live broadcast industry improve the overall technical level of the industry. At the same time, we also welcome industry peers to join the LAS open source community, become practitioners of the LAS standard, and work with Kuaishou to improve the LAS standard.

Official website: https://las-tech.org.cn/

Github: https://github.com/KwaiVideoTeam/las