Loudness Control by Intelligent Audio Content Analysis

Abstract: Automatic audio segmentation aims at extracting information of audio signals. In the case of music tracks, detecting segment boundaries, labelling segments, and detecting repeated segments could be performed. These information can be used in different applications such as creating song summaries and facilitating browsing in music collections. This thesis studies the Foote method which is one of the automatic audio segmentation algorithms. Numerous experiments are carried out to improve the performance of this method. The most suitable parameter settings of the Foote method, which have the best performance in detection of segment boundaries as correct as possible in real time audio segmentation, have been selected. Finally, real time audio segmentation results were applied to automatically control loudness level of a streaming audio input signal. Our experiments show that this application results in better dynamic structure preserving and faster loudness level adjustment.