Abstract
Onset detection is a typical digital signal processing task in acoustic signal analysis, with many applications as in the musical field. Many techniques have been proposed so far, which are typically reliable in terms of performances but often not suitable to real-time computing, for example, they require knowledge of the whole piece to perform optimally, or they are too computationally intensive for most embedded processors. Up to the authors' knowledge, the real-time implementation problem for musical onset detection has been scarcely addressed within the literature, which has motivated them to propose a scalable and computationally efficient algorithm with good detection capabilities. Comparison with other techniques and porting to a real-time embedded processor are discussed as well: provided experimental results seem to confirm the effectiveness of the approach.
Highlights
Onset detection is a fundamental task in many music DSP scenarios
It can be of interest to scale the most recent techniques for onset detection to fit into such platforms
Other methods exist, which make use of neural networks or similar machine-learning structures to analyze the signal features or to combine the extracted features to obtain the final decision. Other papers, such as [5], making use of more efficient DSP techniques, may combine information coming from different onset detection algorithms, requiring an increase in computational cost and a complex decision taking mechanism at the end of the signal processing
Summary
Onset detection is a fundamental task in many music DSP scenarios. Some of these have no critical time constraints, such as database queries, content retrieval, data mining, and so forth. Other methods exist (see, e.g., [3, 4]), which make use of neural networks or similar machine-learning structures to analyze the signal features or to combine the extracted features to obtain the final decision Other papers, such as [5], making use of more efficient DSP techniques, may combine information coming from different onset detection algorithms, requiring an increase in computational cost and a complex decision taking mechanism at the end of the signal processing.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
