The industry alliance "Alliance for Open Media" (AOMedia) recently released version 1.0.0 of the AV2 specification and reference code, marking the first important milestone for this new generation of open source video encoding standards. On this basis, software and hardware manufacturers can now begin to integrate AV2 into various media and technology products.
AV2 is built on the same technology as AV1 and continues the "open source, royalty-free" route. It is mainly targeted at video workloads such as streaming media transmission and transcoding, and provides alternatives to standards such as AVC/H.264, HEVC/H.265 and newer VVC (Versatile Video Coding) that still require the payment of patent fees. The final AV1 specification was released in 2018, and industry adoption has accelerated in the ensuing years, with major streaming platforms including Netflix accelerating deployment of this royalty-free encoding format. In contrast, the AV2 specification was finalized later than originally expected and was originally expected to be launched in 2025. Now when it is officially released, the official positions it as a "next-generation standard" that comprehensively surpasses AV1 in terms of image quality, feature support, and compression efficiency.
According to early test results, AV2’s compression efficiency is about 30% higher than that of AV1, and it has more obvious advantages when facing previous generation encoding standards such as VP9 and H.264. AOMedia said that AV2 is designed to provide "better" compression efficiency, allowing content providers to maintain high-quality video images at a significantly reduced bit rate, or further improve visual effects at the same bit rate. For large-scale streaming media service providers and content distribution networks, this is expected to directly translate into reduced bandwidth costs and improved user perception.
From the perspective of application scenarios, AV2 is optimized for the diverse needs of the current video industry, covering online video on demand and live broadcast, traditional radio and television, and real-time video conferencing. The new standard also strengthens support for mixed reality scenes and takes into account multi-channel video splicing (such as split-screen images showing multiple signals simultaneously) and more levels of image quality options to adapt to different devices and network conditions from mobile terminals to large-screen terminals.
AOMedia was founded in 2015 with the goal of promoting the formulation and implementation of a new generation of open video standards after Google VP9. The members of this non-profit organization include large technology companies such as Amazon, Google, Intel, Nvidia, Microsoft, etc., as well as smaller technology institutions such as Mozilla. Its members span cloud service providers, chip suppliers, browsers and software manufacturers. It aims to reduce the fragmentation and high costs caused by video patent standards through industrial collaboration.
Before AV2, AV1 launched by AOMedia had achieved initial success in the industry. Currently, many modern GPUs from Nvidia and AMD already support AV1 hardware encoding and decoding and are compatible with traditional formats. However, hardware support for VVC, which is also designed for efficient compression, is relatively lacking, although the standard was released as early as 2020. Like AV1, AV2 also hopes to avoid the complex licensing structure of the patent pool through a royalty-free model, further weakening the influence of traditional patented video standards in the market.
In contrast, VVC is subject to complex authorization and charging models at the practical application level, and there are many related patents and licensing entities, which clouds its promotion prospects in the fields of streaming media and consumer hardware. Many industry observers point out that in the current context of rising cost pressures and compliance risks, patent standards with complex licensing structures are more difficult to gain widespread support, while open source royalty-free solutions represented by AV1 and AV2 are more attractive.
Although it may still take many years for AV2 to obtain native hardware acceleration support in GPUs and various terminal SoCs, related deployments have already begun at the software ecosystem level. The open source multimedia project VideoLAN community recently released the portable AV2 decoder dav2d, which performs targeted optimization on different architectures such as x86 (AVX2), ARM (AArch64 NEON), RISC‑V, etc., and provides basic implementation for AV2 decoding on desktops, servers, and embedded devices. The first implementation of this type of software decoder will provide technical support for the pilot deployment of browsers, players and streaming media servers.
On the whole, the official release of AV2 specifications means that the industry continues to move forward on the open source royalty-free video standard, seeking a new balance between compression efficiency, cost control and ecological openness. As more software and hardware manufacturers participate in implementation and optimization, whether AV2 can gain a foothold in the mainstream streaming media and consumer hardware fields like AV1 will become a key observation indicator in the field of video encoding in the next few years.