Optimized Adaptive Loop Filter in Versatile Video Coding

2026-07-08 · Source: cs.CV updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Video Coding · Depth: Expert, long

Summary

An optimized Adaptive Loop Filter (ALF) framework for Versatile Video Coding (VVC) addresses the high coding complexity and extensive picture buffer accesses of the existing ALF, which includes Geometry transformation-based Adaptive Loop Filter (GALF) and Cross Component Adaptive Loop Filter (CCALF). The proposed framework features a parallel design for GALF and CCALF, an adaptive parameter decision scheme for GALF, and a one-pass CCALF method that effectively estimates filtering distortion without conducting filter operations. This optimization reduces picture buffer access from 152 to 1 and achieves roughly 25% time-savings for the ALF module in VTM-8.0 under RA configuration, with negligible coding performance change. Some of these methods have been adopted in the VVC reference software.

Key takeaway

For video encoder developers designing real-time VVC systems, you should integrate the proposed optimized ALF framework. This approach significantly reduces external memory access from 152 to 1 and achieves roughly 25% ALF module time-savings, making high-performance ALF practical for fast presets. Consider adopting parallel GALF/CCALF, adaptive GALF parameter decisions, and the one-pass CCALF scheme to improve efficiency without compromising coding quality.

Key insights

Optimizing VVC's Adaptive Loop Filter significantly reduces encoding complexity and memory access while maintaining coding performance.

Principles

• Parallelize dependent video coding tools.
• Adaptively simplify complex parameter decisions.
• Estimate distortion to avoid costly operations.

Method

The optimized ALF framework involves parallel GALF/CCALF, adaptive GALF parameter decision based on QP/linear GALF, and a one-pass CCALF scheme using distortion estimation.

In practice

• Implement parallel processing for GALF and CCALF.
• Use resolution-dependent models for filter parameter prediction.
• Employ distortion estimation to reduce CCALF passes.

Topics

• Versatile Video Coding
• Adaptive Loop Filter
• Video Compression
• Encoder Optimization
• Memory Access Reduction
• Real-time Encoding

Best for: Research Scientist, AI Scientist, AI Hardware Engineer, Computer Vision Engineer

Related on AIssential

• Blind Quality Enhancement of Compressed Video via Fine-Grained Degradation-Guided Sequential Inference — Fine-grained, multi-scale degradation representation and adaptive processing stages significantly improve blind compressed video quality enhancement.
• Scaling View Synthesis Transformers — Encoder-decoder architectures can be compute-optimal for view synthesis transformers with proper design and training strategies.
• AdaCodec: A Predictive Visual Code for Video MLLMs — AdaCodec uses a predictive visual code to reduce redundancy in video MLLMs, improving efficiency and performance.
• Active Sampling for Ultra-Low-Bit-Rate Video Compression via Conditional Controlled Diffusion — Conditional diffusion models can achieve ultra-low-bit-rate video compression through sparse, adaptive conditioning.
• VMAF v1: Good Is Not Good Enough — VMAF v1 improves video quality assessment by integrating advanced perceptual features and a unified model for diverse viewing conditions.
• Déjà View: Looping Transformers for Multi-View 3D Reconstruction — DéjàView uses a single looped transformer block recurrently for efficient multi-view 3D reconstruction, outperforming larger feed-forward models.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.CV updates on arXiv.org.