Finding the starting time of events (onsets) is useful in a number of applications for audio signals. The goal of this paper is to present a combination of techniques for automatic detection of events in audio signals. The proposed system uses a supervised classification algorithm to combine a set of features extracted from the audio signal and reduce the original signal to a robust detection function. Onsets are obtained by using a simple peak-picking algorithm. This paper describes the analysis system used to extract the features and the details of the neural network algorithm used to combine them. We conclude by comparing the performance of the proposed algorithm with the system that obtained the first place in the 2005 Music Information Retrieval Evaluation eXchange.
(See document for exact affiliation information.)
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member Join the AES. If you need to check your member status, login to the Member Portal.
