publications | Music Informatics Group

2026

Watcharasupat, Karn N.; Lerch, Alexander

Separate This, and All of these Things Around It: Music Source Separation via Hyperellipsoidal Queries Proceedings Article

In: Proceedings of the International Conference on Acoustics, Speech and Signal Processing (ICASSP), Institute of Electrical and Electronics Engineers (IEEE), Barcelona, Spain, 2026.

Abstract | Links | BibTeX | Tags: audio signal processing, audio source separation, Computer Science - Information Retrieval, Computer Science - Machine Learning, Computer Science - Sound, Electrical Engineering and Systems Science - Audio and Speech Processing

@inproceedings{watcharasupat_separate_2026,

title = {Separate This, and All of these Things Around It: Music Source Separation via Hyperellipsoidal Queries},

author = {Karn N. Watcharasupat and Alexander Lerch},

url = {http://arxiv.org/abs/2501.16171},

doi = {10.48550/arXiv.2501.16171},

year  = {2026},

date = {2026-01-01},

urldate = {2026-01-01},

booktitle = {Proceedings of the International Conference on Acoustics, Speech and Signal Processing (ICASSP)},

publisher = {Institute of Electrical and Electronics Engineers (IEEE)},

address = {Barcelona, Spain},

abstract = {Music source separation is an audio-to-audio retrieval task of extracting one or more constituent components, or composites thereof, from a musical audio mixture. Each of these constituent components is often referred to as a "stem" in literature. Historically, music source separation has been dominated by a stem-based paradigm, leading to most state-of-the-art systems being either a collection of single-stem extraction models, or a tightly coupled system with a fixed, difficult-to-modify, set of supported stems. Combined with the limited data availability, advances in music source separation have thus been mostly limited to the "VDBO" set of stems: textbackslashtextitvocals, textbackslashtextitdrum, textbackslashtextitbass, and the catch-all textbackslashtextitothers. Recent work in music source separation has begun to challenge the fixed-stem paradigm, moving towards models able to extract any musical sound as long as this target type of sound could be specified to the model as an additional query input. We generalize this idea to a textbackslashtextitquery-by-region source separation system, specifying the target based on the query regardless of how many sound sources or which sound classes are contained within it. To do so, we propose the use of hyperellipsoidal regions as queries to allow for an intuitive yet easily parametrizable approach to specifying both the target (location) as well as its spread. Evaluation of the proposed system on the MoisesDB dataset demonstrated state-of-the-art performance of the proposed system both in terms of signal-to-noise ratios and retrieval metrics.},

keywords = {audio signal processing, audio source separation, Computer Science - Information Retrieval, Computer Science - Machine Learning, Computer Science - Sound, Electrical Engineering and Systems Science - Audio and Speech Processing},

pubstate = {published},

tppubtype = {inproceedings}

}

2023

Lerch, Alexander

An Introduction to Audio Content Analysis: Music Information Retrieval Tasks and Applications Book

2, Wiley-IEEE Press, Hoboken, N.J, 2023, ISBN: 978-1-119-89094-2.

Abstract | Links | BibTeX | Tags: analysis, audio, Audio content analysis, audio signal processing, Automatic Music Transcription, Computer sound processing, machine listening, Matlab, MIR, music analysis, music informatics, music information retrieval, Python

@book{lerch_introduction_2023,

title = {An Introduction to Audio Content Analysis: Music Information Retrieval Tasks and Applications},

author = {Alexander Lerch},

url = {https://ieeexplore.ieee.org/servlet/opac?bknumber=9965970},

isbn = {978-1-119-89094-2},

year  = {2023},

date = {2023-01-01},

urldate = {2022-01-01},

publisher = {Wiley-IEEE Press},

address = {Hoboken, N.J},

edition = {2},

abstract = {An Introduction to Audio Content Analysis Enables readers to understand the algorithmic analysis of musical audio signals with AI-driven approaches An Introduction to Audio Content Analysis serves as a comprehensive guide on audio content analysis explaining how signal processing and machine learning approaches can be utilized for the extraction of musical content from audio. It gives readers the algorithmic understanding to teach a computer to interpret music signals and thus allows for the design of tools for interacting with music. The work ties together topics from audio signal processing and machine learning, showing how to use audio content analysis to pick up musical characteristics automatically. A multitude of audio content analysis tasks related to the extraction of tonal, temporal, timbral, and intensity-related characteristics of the music signal are presented. Each task is introduced from both a musical and a technical perspective, detailing the algorithmic approach as well as providing practical guidance on implementation details and evaluation. To aid in reader comprehension, each task description begins with a short introduction to the most important musical and perceptual characteristics of the covered topic, followed by a detailed algorithmic model and its evaluation, and concluded with questions and exercises. For the interested reader, updated supplemental materials are provided via an accompanying website. Written by a well-known expert in the music industry, sample topics covered in Introduction to Audio Content Analysis include: Digital audio signals and their representation, common time-frequency transforms, audio features Pitch and fundamental frequency detection, key and chord Representation of dynamics in music and intensity-related features Beat histograms, onset and tempo detection, beat histograms, and detection of structure in music, and sequence alignment Audio fingerprinting, musical genre, mood, and instrument classification An invaluable guide for newcomers to audio signal processing and industry experts alike, An Introduction to Audio Content Analysis covers a wide range of introductory topics pertaining to music information retrieval and machine listening, allowing students and researchers to quickly gain core holistic knowledge in audio analysis and dig deeper into specific aspects of the field with the help of a large amount of references.},

keywords = {analysis, audio, Audio content analysis, audio signal processing, Automatic Music Transcription, Computer sound processing, machine listening, Matlab, MIR, music analysis, music informatics, music information retrieval, Python},

pubstate = {published},

tppubtype = {book}

}

An Introduction to Audio Content Analysis Enables readers to understand the algorithmic analysis of musical audio signals with AI-driven approaches An Introduction to Audio Content Analysis serves as a comprehensive guide on audio content analysis explaining how signal processing and machine learning approaches can be utilized for the extraction of musical content from audio. It gives readers the algorithmic understanding to teach a computer to interpret music signals and thus allows for the design of tools for interacting with music. The work ties together topics from audio signal processing and machine learning, showing how to use audio content analysis to pick up musical characteristics automatically. A multitude of audio content analysis tasks related to the extraction of tonal, temporal, timbral, and intensity-related characteristics of the music signal are presented. Each task is introduced from both a musical and a technical perspective, detailing the algorithmic approach as well as providing practical guidance on implementation details and evaluation. To aid in reader comprehension, each task description begins with a short introduction to the most important musical and perceptual characteristics of the covered topic, followed by a detailed algorithmic model and its evaluation, and concluded with questions and exercises. For the interested reader, updated supplemental materials are provided via an accompanying website. Written by a well-known expert in the music industry, sample topics covered in Introduction to Audio Content Analysis include: Digital audio signals and their representation, common time-frequency transforms, audio features Pitch and fundamental frequency detection, key and chord Representation of dynamics in music and intensity-related features Beat histograms, onset and tempo detection, beat histograms, and detection of structure in music, and sequence alignment Audio fingerprinting, musical genre, mood, and instrument classification An invaluable guide for newcomers to audio signal processing and industry experts alike, An Introduction to Audio Content Analysis covers a wide range of introductory topics pertaining to music information retrieval and machine listening, allowing students and researchers to quickly gain core holistic knowledge in audio analysis and dig deeper into specific aspects of the field with the help of a large amount of references.

2012

Lerch, Alexander

An Introduction to Audio Content Analysis: Applications in Signal Processing and Music Informatics Book

Wiley-IEEE Press, Hoboken, 2012, ISBN: 978-1-118-26682-3.

Abstract | Links | BibTeX | Tags: analysis, audio, audio signal processing, information, listening, machine, machine listening, music, music analysis, music information retrieval, processing, retrieval, signal

Music Informatics Group