Block-Partitioning Strategies for Accelerated Multi-rate Encoding in Adaptive VVC Streaming

2025-10-16

AI Summary

High computational complexity of multi-rate VVC encoding hinders practical deployment in adaptive streaming. Method: This paper proposes a fast multi-QP encoding method based on CU depth information reuse: the single- or double-boundary depth constraints derived from a reference QP guide CU partitioning decisions for other QPs, significantly reducing redundant computations. Integrated into the VVenC encoder, the method supports all standard presets without modifying bitstream syntax or structure. Results: Experimental evaluation on mainstream test sequences shows an average 11.69% reduction in encoding time, with negligible bitrate overhead (<0.6%). The approach achieves Pareto-optimal improvements in the BD-rate versus encoding time trade-off, substantially enhancing both the practicality and scalability of multi-rate VVC encoding for adaptive streaming applications.

Technology Category

• Computer Vision: Multi-modal Vision
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

The demand for efficient multi-rate encoding techniques has surged with the increasing prevalence of ultra-high-definition (UHD) video content, particularly in adaptive streaming scenarios where a single video must be encoded at multiple bitrates to accommodate diverse network conditions. While Versatile Video Coding (VVC) significantly improves compression efficiency, it introduces considerable computational complexity, making multi-rate encoding a resource-intensive task. This paper examines coding unit (CU) partitioning strategies to minimize redundant computations in VVC while preserving high video quality. We propose single- and double-bound approaches, leveraging CU depth constraints from reference encodes to guide dependent encodes across multiple QPs. These methods are evaluated using VVenC with various presets, demonstrating consistent improvements in encoding efficiency. Our methods achieve up to 11.69 % reduction in encoding time with minimal bitrate overhead (<0.6 %). Comparative Pareto-front (PF) analysis highlights the superior performance of multi-rate approaches over existing configurations. These findings validate the potential of CU-guided strategies for scalable multi-rate encoding in adaptive streaming.

Problem

• Reducing computational complexity in VVC multi-rate encoding
• Minimizing redundant computations while preserving video quality
• Optimizing CU partitioning strategies for adaptive streaming efficiency

Innovation

• CU depth constraints guide dependent encodes
• Single- and double-bound approaches reduce computations
• Partitioning strategies minimize redundant VVC encoding