Compression Efficiency and Delay Tradeoffs for Hierarchical B-Pictures and Pulsed-Quality Frames

Abstract

Real-time video applications require tight bounds on end-to-end delay. Hierarchical bidirectional prediction requires buffering frames in the encoder input buffer, thereby contributing to encoder input delay. Long-term frame prediction with pulsed quality requires buffering at the encoder output, increasing the output buffer delay. Both hierarchical B-pictures and pulsed-quality coders involve uneven bit-rate allocation. Both the encoder and decoder buffering requirements depend on the rate allocation. We derive an efficient rate allocation for hierarchical B-pictures using the power spectral density of a wide-sense stationary process. In addition, we discuss important aspects of hierarchical predictive coding, such as the effect of the temporal prediction distance and delay tradeoffs for prediction branch truncation. Finally, we investigate experimentally the tradeoff between delay and compression efficiency.

Error resilience video coding parameters and mechanisms selection with End-to-End rate-distortion analysis at frame level

“…R primary and R red can be estimated using the RD model [4,8]. The variance of the prediction error is formulated with the power spectral density of the prediction error [4].…”

“…The variance of the prediction error is formulated with the power spectral density of the prediction error [4]. In [8], the Fig. 4 The actual and average D dif of different sequences, when the redundant picture uses different delta quantization parameters ΔQp prediction error σ Δ is then parameterized with the temporal prediction distance Δt by setting σ Δ =σ Δ (Δt).…”

“…This formula is statistically trained for the estimation of bit allocation of different temporal levels at high bit rate [8]. The statistical relation between the temporal reference distance and the bit rate has also been studied through our experiments as illustrated in Fig.…”

“…The encoding parameter should be adjusted according to the source distortion propagation. In the rate distortion function [8], the D q,i of f(i) can be formulated as: where D mcp,i denotes motion compensation distortion (MCP), R i denotes the bitrate, i denotes the frame index, and b denotes a constant related to the source distribution. Decoupling D mcp,i , it can be represented as α×(D q,i-1 + D omcp,i ) where D omcp,i denotes the MCP distortion of the original ith frame, and α is equal to 0.94 from experiment [22].…”

“…Assume that the temporal prediction distance of the current frame is Δt. D omcp,i can be taken as the statistical function g(Δt) of Δt [8]. In previous research, the influenced D mcp,i can be calculated from (25):…”