Coding Tree Unit

Description

A Coding Tree Unit (CTU) is the basic processing element in modern video compression standards, specifically High Efficiency Video Coding (HEVC) and Versatile Video Coding (VVC). It represents a square block of luma samples and corresponding chroma samples, with typical sizes ranging from 64×64 to 128×128 pixels. The CTU serves as the starting point for the recursive quadtree partitioning process that defines the coding structure of the entire frame.

Each CTU can be partitioned into smaller Coding Units (CUs) using a quadtree decomposition. This partitioning is adaptive and determined during the encoding process based on rate-distortion optimization. The CTU contains not only the pixel data but also the associated syntax elements that describe how it should be partitioned, predicted, and transformed. This hierarchical structure allows the codec to efficiently handle varying levels of detail within different regions of the video frame.

Within each CTU, the partitioning creates a tree structure where each leaf node represents a Coding Unit. These CUs can then be further divided into Prediction Units (PUs) for motion compensation and Transform Units (TUs) for spatial transformation. The CTU concept replaces the macroblock structure used in earlier standards like H.264/AVC, providing greater flexibility in block size selection and improved compression efficiency for high-resolution video content.

The size of the CTU is a critical parameter that affects both compression performance and computational complexity. Larger CTUs (like 128×128) generally provide better compression for homogeneous regions but require more processing power. The CTU structure enables parallel processing opportunities since multiple CTUs can be processed independently, facilitating hardware acceleration and multi-core implementations. This makes CTUs essential for real-time encoding and decoding of ultra-high-definition video content.

In the 3GPP context, CTUs are particularly relevant for video services over mobile networks, where efficient compression directly impacts bandwidth utilization and quality of experience. The CTU-based coding structure allows for better adaptation to network conditions through scalable video coding and error resilience techniques. This hierarchical approach also supports advanced features like screen content coding and adaptive resolution changes within the same video stream.

Purpose & Motivation

The CTU was introduced to address the limitations of the fixed-size macroblock structure used in previous video coding standards like H.264/AVC. As video resolutions increased from standard definition to 4K and 8K, the 16×16 macroblock became inefficient for representing large homogeneous areas and insufficient for capturing fine details in complex textures. The CTU's larger size and hierarchical partitioning provide the necessary flexibility to optimize compression for modern high-resolution video content.

Traditional macroblock-based coding struggled with compression efficiency for ultra-high-definition content because it couldn't adapt block sizes effectively across varying spatial characteristics. The CTU's quadtree structure enables the codec to use larger blocks for smooth areas (reducing overhead) and smaller blocks for detailed regions (preserving quality). This adaptive approach significantly improves compression ratios while maintaining visual quality, which is crucial for bandwidth-constrained mobile networks.

The development of CTU-based coding was motivated by the growing demand for high-quality video services over 3GPP networks. As mobile video consumption increased exponentially, network operators needed more efficient compression to deliver 4K streaming, video conferencing, and immersive media within available bandwidth. The CTU architecture also facilitates hardware-friendly implementations with better parallel processing capabilities, enabling real-time encoding and decoding on mobile devices with limited computational resources.