AV1

AV1 (AOMedia Video 1) is a royalty-free video coding format designed for internet video delivery, standardized by the Alliance for Open Media (AOMedia) in 2018. Unlike its predecessors H.264 and HEVC, which were developed within traditional standards bodies and burdened by patent pool licensing, AV1 emerged from a consortium of technology companies — Google, Mozilla, Microsoft, Amazon, Netflix, and others — united by a shared interest in reducing the royalty tax on streaming infrastructure. AV1 represents not merely a technical advance in compression efficiency but a structural shift in how video standards are produced, governed, and distributed.

Technical Architecture

AV1 achieves roughly 30% better compression than H.264 and comparable efficiency to HEVC, but without the patent licensing overhead. Its technical innovations include:

Expanded transform sizes: While JPEG and MPEG-2 fixed transform blocks at 8×8 pixels and H.264 extended this modestly, AV1 supports transforms up to 64×64, adapting to the scale of features within each frame. This is not merely an incremental improvement — it reflects a changed assumption about display resolution. The 8×8 block was designed for standard-definition television; the 64×64 block assumes 4K and 8K displays where spatial correlations extend across larger regions.

Constrained directional enhancement filters: AV1 employs in-loop filters that reduce blocking and ringing artifacts without the computational cost of H.264's adaptive deblocking. The filter design encodes a perceptual model developed through large-scale subjective testing — a rare instance of empirical human vision research directly shaping codec design.

Tile-based parallelism: Recognizing that parallel computing had become essential for real-time video encoding, AV1 structures frames into independently decodable tiles. This architectural decision acknowledges the memory wall and power wall constraints that ended the era of single-threaded performance growth. The tile structure is a concession to physics: when you cannot make one processor faster, you must make the problem more parallel.

Film grain synthesis: Perhaps AV1's most unusual feature is the ability to separate film grain from image content, transmit the grain as metadata, and synthesize it at the decoder. This is a radical departure from the lossy compression paradigm of discarding imperceptible information. Instead, AV1 recognizes that film grain is perceptually significant but informationally redundant — and treats these two properties separately. The image is compressed conventionally; the grain is synthesized. This separation of perceptual property from information content hints at a deeper truth about compression: that what matters to human perception is not always what contains the most bits.

The Political Economy of Open Standards

AV1's development cannot be understood without examining the licensing crisis that preceded it. H.264's patent pool, administered by MPEG LA, imposed per-unit royalties that became a significant cost for high-volume streaming services. When Google acquired On2 Technologies and open-sourced VP8 as WebM in 2010, it was a direct challenge to this model — but VP8's technical performance lagged H.264, and patent threats from MPEG LA licensees limited adoption.

HEVC, standardized in 2013, promised better compression but multiplied the licensing problem. Three competing patent pools emerged (MPEG LA, HEVC Advance/Access Advance, and Velos Media), with overlapping and conflicting claims. The total royalty burden exceeded what many streaming services could bear. The licensing fragmentation was not an accident but a structural feature of standards-development organizations that permit participants to declare essential patents after standardization is complete.

AV1 was designed explicitly to escape this trap. AOMedia required participants to license their contributions on royalty-free terms as a condition of membership. The defensive patent clause ensures that if a member sues another member over AV1, their own license terminates. This is not merely a technical standard — it is a legal architecture designed to solve a coordination problem that traditional standards bodies could not.

AV1 as Systems Intervention

From a systems perspective, AV1 is fascinating because it intervenes at multiple levels simultaneously. At the technical level, it improves compression efficiency. At the economic level, it eliminates licensing friction. At the political level, it realigns incentives around open collaboration rather than patent extraction. At the organizational level, it demonstrates that industry consortia can produce standards faster and with broader participation than traditional bodies like ITU-T and ISO/IEC.

The result is a compression standard that is not merely better but differently structured. H.264 and HEVC were designed by committees optimized for patent incorporation; AV1 was designed by a consortium optimized for patent exclusion. The technical differences — transform sizes, filter choices, entropy coding methods — are inseparable from these organizational differences. A codec designed to avoid patents makes different tradeoffs than a codec designed to accommodate them.

AV1 is often framed as a technical response to H.264 and HEVC — a better codec that happens to be free. This framing inverts the causality. AV1 exists because the licensing structure of previous standards created a market failure that no single company could solve alone. The compression efficiency is real, but it is secondary. The real achievement of AV1 is proving that open standards can emerge from industrial coordination without state sponsorship or academic mediation — and that when they do, they can outcompete standards produced by institutions with a century of accumulated legitimacy. The question is not whether AV1 will displace HEVC. The question is what other coordination failures in technology standards might be solved by the same mechanism.