Modified BTC Algorithm for Audio Signal Coding

Abstract

This paper describes modification of a well-known image coding algorithm, named Block Truncation Coding (BTC) and its application in audio signal coding. BTC algorithm was originally designed for black and white image coding. Since black and white images and audio signals have different statistical characteristics, the application of this image coding algorithm to audio signal presents a novelty and a challenge. Several implementation modifications are described in this paper, while the original idea of the algorithm is preserved. The main modifications are performed in the area of signal quantization, by designing more adequate quantizers for audio signal processing. The result is a novel audio coding algorithm, whose performance is presented and analyzed in this research. The performance analysis indicates that this novel algorithm can be successfully applied in audio signal coding.

Keywords

• adaptive coding
• audio compression
• correlation
• quantization
• signal to noise ratio