Signal Processing, Discrete Autocorrelation Functions, Acoustic Signals, Statistical Significance, Paired Tests, Hypothesis Testing

For purposes of signal analysis, a wide spectrum of methods has appeared in the mathematical statistics. With regards to a random behavior of considered signals, the methods are based on the probability theory. In the signal processing theory, the autocorrelation functions can be considered as a suitable tool for an identification of a tightness of bindings within an analyzed signal. However, this type of analysis is based only on the descriptive approach. A comparison of autocorrelation functions based on principles of the testing hypotheses can be advantageous for an evaluation of a tightness of bindings across more than one analyzed signal. This type of comparison of signal properties has not been so widely presented yet. In favor of this aim, a proposal of a comparison of the estimated autocorrelation functions is presented in this paper. The real acoustic signals were recorded in the Campus studio of RadioIgel at University College of Teacher Education Styria. Then these signals were modified using particular types of sound effects. The proposed analysis is presented for a selected part of the acoustic signals before and after modifications.