Fractal Audio Coding

We explore the performance of applying fractal coding on audio data. Some conventional fractal coding problems have been studied with audio data to provide an overview on this subject. A review of fractal coding is presented. We implement a fractal audio coding scheme to carry out the experiments. The performance of the scheme can be controlled by a number of different parameters, including mapping tolerance, scaling factor, partition range, domain specification, and bit allocation. Empirical results have been obtained from experimenting on various audio data in our testing set under different parameter combinations. Some conclusions and suggestions have been made by analyzing and comparing experimental results. The study leads us to conclude that fractal coding is not an appropriate model to be applied alone to complex audio data. A major barrier is the inability to represent smooth continuous functions. With the new trend of integrating some other methods in fractal coding research, we discuss some future aspects of fractal audio coding, and some possible improvements by combining it with other techniques such as wavelet transforms.