Presently reported is a wavelet-based method for the temporal localization of sudden onsets in audio signals with sub-millisecond precision. The method only requires O(n) operations, which is highly efficient. The entire audio signal can be processed as a whole without the need to be broken down into individual windowed overlapping blocks. It can also be processed in a streaming mode compatible with real-time processing. In comparison with time-domain and frequency-domain methods, the wavelet-based method proposed here offers several distinct advantages in sudden onset detection, temporal localization accuracy, and computational cost, which may therefore find broad applications in audio signal processing and music information retrieval.
Wan, Yuxuan; Chen, Yijia; Sim, Keegan Yi Hang; Wu, Lijia; Geng, Xianzheng; Chau, Kevin
Affiliation: Hong Kong University of Science and Technology, Clean Water Bay, Hong Kong
AES Convention: 147 (October 2019) Paper Number: 10280
Publication Date: October 8, 2019
Subject: Semantic Audio
Download Now (676 KB)
No AES members have commented on this paper yet.
If you are not yet an AES member and have something important to say about this paper then we urge you to join the AES today and make your voice heard. You can join online today by clicking here.