Versatile Video Coding

Description

Versatile Video Coding (VVC), standardized as H.266 by ITU-T and MPEG-I Part 3 (ISO/IEC 23090-3), is the latest generation video codec incorporated into the 3GPP specifications for multimedia services. Its primary function is to compress digital video with significantly higher efficiency than previous standards like High Efficiency Video Coding (HEVC/H.265) and Advanced Video Coding (AVC/H.264). 3GPP has standardized the use of VVC for media delivery over 5G networks, defining profiles, levels, and carriage formats in specifications such as 26.927 and 26.928 to ensure interoperability between servers, networks, and user equipment.

Architecturally, VVC builds upon the hybrid block-based coding model but introduces numerous advanced tools that contribute to its efficiency. The core unit is the Coding Tree Unit (CTU), which can be partitioned using a sophisticated multi-type tree structure allowing binary, ternary, and quad-tree splits, enabling a much more precise adaptation to video content. Key technical components include enhanced intra-prediction with 93 directional modes (compared to 33 in HEVC), affine motion prediction for complex non-translational motion, adaptive loop filters (ALF) to reduce artifacts, and decoder-side motion vector refinement. For immersive media, VVC introduces specific tools for 360-degree video, such as equirectangular projection and region-wise packing. The codec operates by dividing a video sequence into slices and tiles for parallel processing, applying spatial (intra) and temporal (inter) prediction, transforming residual data, and applying entropy coding (CABAC).

Within the 3GPP ecosystem, VVC is integrated into the Multimedia Broadcast and Multicast Service (MBMS) and Packet Switched Streaming Service (PSS). Its role is to enable bandwidth-efficient delivery of demanding video applications over 5G New Radio (NR). This is critical for services like 4K/8K broadcasting, augmented reality (AR), virtual reality (VR), and cloud gaming, where high resolution, high frame rates, and low latency are paramount. 3GPP specifications define how VVC bitstreams are packetized for transport over RTP/IP or MPEG-2 TS, and how Dynamic Adaptive Streaming over HTTP (DASH) manifests signal VVC capabilities. The network's Quality of Service (QoS) framework can be applied to VVC media flows to guarantee the necessary bitrate and packet loss performance, making it a fundamental enabler for the enhanced Mobile Broadband (eMBB) and massive Internet of Things (mIoT) service categories that include rich media.

Purpose & Motivation

VVC was created to address the exponential growth in video traffic, which dominates mobile networks, and the emergence of new, data-intensive video formats. Its predecessor, HEVC, while efficient, faced licensing complexities and was reaching its compression limits for next-generation applications like 8K resolution (up to 7680x4320 pixels), High Dynamic Range (HDR) video, and immersive 360-degree content. Streaming such content with HEVC would consume prohibitive amounts of scarce wireless spectrum, increasing costs and degrading network performance for all users.

The motivation for standardizing VVC within 3GPP, starting in Release 16, was to future-proof 5G networks for the media landscape of the 2020s and beyond. 5G promises multi-gigabit data rates and ultra-low latency, but without a corresponding leap in compression efficiency, these capabilities would be wasted merely carrying overhead. VVC's approximate 50% bitrate reduction for the same visual quality as HEVC directly translates to serving twice as many users with high-quality video or enabling previously impractical services like live 8K broadcasting to mobile devices. Furthermore, by standardizing its use, 3GPP ensures a consistent, interoperable media layer across global 5G deployments, avoiding fragmentation and encouraging widespread adoption by device and service providers.

Key Features

• Approximately 50% better compression efficiency than HEVC for equivalent video quality
• Support for ultra-high-definition video resolutions up to 8K and beyond
• Advanced partitioning structures (multi-type tree) for precise block segmentation
• Enhanced prediction tools including affine motion and 93 intra-prediction modes
• Specific coding tools for 360-degree omnidirectional immersive video
• Support for High Dynamic Range (HDR) and Wide Color Gamut (WCG) content

Evolution Across Releases

Defining Specifications