Method of selecting a requantization step size and controlling a bit-rate

Abstract

A method selects a requantization step size which abruptly changes a bit-rate in a piecewise linearly decreasing model and controls the bit-rate. The method includes: deciding a minimum value of the requantization step size Q2, satisfying 2·Q1+nint(p·Q2/q)<2·Q2 for a quantization step size Q1 of a coded image block of a video image, as requantization step size information; determining a minimum value of a requantization step size Q2, satisfying 4·Q1+nint(p·Q2/q)<2·Q2 for the quantization step size Q1 of the coded image block, as the requantization step size information; taking a minimum value of a requantization step size Q2, satisfying 3·Q1<2·Q2 for the quantization step size Q1 of the coded image block, as the requantization step size information; and choosing a minimum value of a requantization step size Q2, satisfying 5·Q1<2·Q2 for the quantization step size Q1 of the coded image block, as the requantization step size information, wherein nint(·) is a nearest integer operator and p and q represent positive integers, respectively.

Dependencies are inferred from the claim text.