Efficient Modeling and Low Complexity Implementation of Rate Estimation in Versatile Video Coding

Abstract

In Versatile Video Coding (VVC), Rate-Distortion Optimized Quantization (RDOQ) is a widely adopted technique to strike a balance between bit rate and distortion. However, the computational complexity introduced by RDOQ poses significant challenges for real-time applications. To address this issue, we propose a low-complexity rate estimation model to accelerate the conventional RDOQ approach. Through comprehensive analysis, we have identified the generalized Gaussian distribution (GGD) as an efficient model for simulating the distribution of transform coefficients. Leveraging this insight, we developed a GGD-based rate estimator to expedite the selection of quantization levels. Experimental results demonstrate that the proposed model achieves an average encoding time reduction of 13.15%, with only a minimal impact on BD-Rate.