A Performance Evaluation of Position Encoding as an Improved Cavlc Encoding Method in H.264 Video Compression Standard

Context Adaptive Variable Length Coding (CAVLC) is one of the widely used entropy encoding algorithms in Video Compression standards especially in H.264/AVC. The conventional CAVLC generates a binary sequence from the Transformed and Quantized coefficients; this binary code consists of five parts namely, Total Coefficients, Trailing ones, Levels, Run Before and Total Zeros. A new method, Position encoding replaces the last two steps, Run Before and Total Zeros, with a single Position code thus reducing the total code length and increased bit compression along with efficiency. This paper evaluates and extends Position encoding, which is currently limited to 8 coefficients, to 9 coefficients by constructing a new look up table and testing the efficiency in bit compression and computational speed with that of CAVLC standard. It was found that, as the number of coefficients increases, the bit compression reduces although the computational speeds in both cases remain the same. It is undoubtedly more beneficial in terms of less computational complexity, more speed and less hardware, compared to the conventional CAVLC encoding method.