Rate-Distortion Optimization Using Adaptive Lagrange Multipliers

Abstract

In current standardized hybrid video encoders, the Lagrange multiplier determination model is a key component in rate-distortion optimization. This originated some 20 years ago based on an entropy-constrained high-rate approximation and experimental results obtained using an H.263 reference encoder on limited test material. In this paper, we present a comprehensive analysis of the results of a Lagrange multiplier selection experiment conducted on various video content using H.264/AVC and HEVC reference encoders. These results show that the original Lagrange multiplier selection methods, employed in both video encoders, are able to achieve optimum rate-distortion performance for I and P frames, but fail to perform well for B frames. The relationship is identified between the optimum Lagrange multipliers for B frames and distortion information obtained from the experimental results, leading to a novel Lagrange multiplier determination approach. The proposed method adaptively predicts the optimum Lagrange multiplier for B frames based on the distortion statistics of recent reconstructed frames. After integration into both H.264/AVC and HEVC reference encoders, this approach was evaluated on 36 test sequences with various resolutions and differing content types. The results show consistent bitrate savings for various hierarchical B frame configurations with minimal additional complexity. BD savings average approximately 3% when constant QP values are used for all frames, and 0.5% when non-zero QP offset values are employed for different B frame hierarchical levels.

Video compression with low complexity CNN-based spatial resolution adaptation

“…On the other hand, too much high frequency information may lead to higher bitrates during compression. This results in an optimisation problem which is similar to traditional rate distortion optimisation 18,19 in image and video coding. To solve this problem, the ideal solution would be to conduct an end-to-end optimisation which includes an image or video codec in the training loop -to encode the low resolution CNN output and generate a real bitstream whose corresponding bitrate (R) could be measured.…”

Sub-Frame Layer Rate Distortion Optimization for High Efficiency Video Coding

“…Taking (28) into (32), the impact fact of each frame layer and each sub-frame layer is obtained. Given λ g , the optimal λ of frame p i and its sub-frames can be obtained by taking I F p i into (31).…”

“…III-A 2: Pre-encode first WinLen frames in the sequence and collects inter-frame dependency of frames and sub-frames according to Sec. II 3: Calculate the net total impact of each frame layer and each sub-frame layer according to (28) 4: Calculate the global Lagrange multiplier λ g according to (33) 5: Calculate impact factor of each frame layer and impact factor of L0 by taking (28) and (33) into 32 k = k + 1, go to next frame 32: end while sub-frames and LC sub-frames according to Sec. III-A.…”

“…II. After pre-encoding, the net total impact of each frame layer and each sub-frame layer are obtained according to (28). With the net total impact of L4, the global Lagrange multiplier λ g is estimated according to (33).…”

“…In [26], λ-QP mismatch for static background was corrected by exponential regression function. Zhang and Bull [27], [28] modified λ for B frames based on the distortion statistics of recent reconstructed frames. Li et al [29] used a method similar to [4] to explore λ-QP relationship and obtained more HEVC compliant λ-QP function.…”