Joint Video Exploratory Group

Description

The Joint Video Exploratory Group (JVET) was a collaborative technical body formed jointly by the ITU-T Study Group 16 (VCEG) and ISO/IEC JTC 1/SC 29 (MPEG). Within the 3GPP context, JVET's work is referenced and adopted to standardize advanced video coding for multimedia services. The group's primary output was the Versatile Video Coding (VVC) standard, also known as H.266. The architecture of VVC builds upon the hybrid block-based coding framework of its predecessors (like HEVC/H.265) but introduces numerous technical enhancements. It works by dividing a video frame into blocks and using sophisticated prediction (intra-frame and inter-frame), transformation, quantization, and entropy coding techniques to remove spatial and temporal redundancy. Key innovations include a more flexible block partitioning structure using QuadTree plus Multi-Type Tree (QTMT), advanced motion vector prediction, and adaptive loop filters.

The coding process begins with partitioning a picture into Coding Tree Units (CTUs), which can be split recursively using a combination of quadtree and binary/ternary splits, allowing for a much more precise adaptation to the content's detail than the rigid quadtree of HEVC. For prediction, intra-prediction uses a wide array of directional modes, while inter-prediction incorporates advanced techniques like affine motion prediction for complex motions and decoder-side motion vector refinement. The transform stage can apply multiple transform selections (DCT-II, DST-VII) adaptively. After quantization, context-adaptive binary arithmetic coding (CABAC) is used for entropy coding, with several improvements for higher throughput. These components work together to achieve the core objective: a substantial bitrate reduction—approximately 50% over HEVC at the same subjective video quality.

In the 3GPP ecosystem, the specifications (e.g., TS 26.855, TS 26.928) define how VVC is packaged and transported for multimedia services, particularly for 5G. Its role is to enable efficient delivery of bandwidth-intensive video applications such as 4K/8K streaming, 360-degree video, and immersive media over mobile networks. By drastically improving compression efficiency, VVC reduces the network load and storage requirements, which is critical for making advanced video services feasible and economical in capacity-constrained radio access scenarios. 3GPP standards ensure interoperability by specifying codec profiles, levels, and the encapsulation of VVC bitstreams within multimedia containers for session establishment and delivery.

Purpose & Motivation

JVET was established to address the escalating demand for higher video resolutions (4K, 8K, HDR), immersive formats (VR/360°), and more efficient video compression for streaming and storage. The previous standard, High Efficiency Video Coding (HEVC/H.265), while a significant advance over H.264/AVC, was facing challenges including complex patent licensing, and its compression gains were becoming insufficient for the next wave of ultra-high-definition content. The exponential growth of video traffic, projected to dominate mobile data, created an urgent need for a new codec that could dramatically reduce bitrates without sacrificing quality to make future video services sustainable for network operators.

The creation of JVET and the VVC standard was motivated by the desire to achieve a generational leap in coding efficiency. By jointly leveraging the expertise of both ITU-T and ISO/IEC, the group aimed to create a technically superior and widely accepted standard. For 3GPP and the mobile industry, the purpose was clear: to define the video codec that would underpin the 5G multimedia experience. Efficient video coding is a cornerstone for 5G use cases like enhanced Mobile Broadband (eMBB) and massive IoT with video sensors. VVC solves the problem of transporting extremely high-quality video over wireless links where spectrum is a scarce and expensive resource, enabling new business models and user experiences while managing network congestion.

Key Features

• ~50% bitrate reduction compared to HEVC at equivalent visual quality
• Flexible block partitioning with QuadTree plus Multi-Type Tree (QTMT) structure
• Advanced motion compensation with affine motion and decoder-side motion vector refinement
• Enhanced intra-prediction with 67 directional modes
• Adaptive multiple transform selection (DCT-II, DST-VII)
• Improved entropy coding using CABAC with increased throughput

Evolution Across Releases

Defining Specifications