Modelling and Analysing Sound and Music Signals - ELE00189M

«Back to module search

Module summary

This module explores how sound, audio systems and musical signals can be modelled and analysed. Scientific and engineering research and the demands of industry and commercial applications are considered and how they have resulted in a range of approaches to the design, modelling and implementation of contemporary sound models and systems. The consideration of such approaches extends to the production, perception and acoustic features of different sound and music signals and how they might be analysed using suitable frameworks and metrics.

Module will run

Occurrence	Teaching period
A	Semester 1 2025-26

Module aims

Subject content aims:

• To develop knowledge and understanding in the use of models and metrics used to create, process and analyse sound and music signals.

• To introduce the theory supporting a range of models and metrics used to create, process and analyse sound and music signals and their application for varied sound synthesis, audio processing and sound and music signals analysis tasks.

• To develop practical experience in designing digital realisations of a range of sound models and systems.

• To develop practical experience in designing/implementing sound and music analysis metrics to extract features from audio recordings and music content.

• To understand and appreciate recent research and industry practice in these areas and the application of this work across other fields and disciplines.

Graduate skills aims:

• To develop skills in critically evaluating and synthesising new information based on researched information and writing concise technical reports appropriate for the target audience.

Module learning outcomes

Subject content learning outcomes

After successful completion of this module, students will be able to:

• Discuss the theory behind the models and metrics used to create, process and analyse sound and music signals.

• Investigate the application and understanding of a range of models and metrics used to create, process and analyse sound and music signals

• Evaluate, compare and critique algorithmic approaches used in a range of models and metrics used to create, process and analyse sound and music signals

• Implement sound models and systems for creating and processing audio content.

• Implement sound and music analysis metrics to extract relevant features from audio recordings

• Design and create/process a range of sound examples using one or more sound model/system implementations to demonstrate their capabilities and limitations.

• Evaluate and interpret the effectiveness of sound and music analysis metrics to extract relevant features from audio recordings.

• Interpret and communicate theoretical and practical examples of models and metrics used to create, process and analyse sound and music signals through publication.

Graduate skills learning outcomes

After successful completion of this module, students will be able to:

• Construct concise technical reports that critically evaluate and synthesise new information based on research, appropriate for the target audience

• Design, deliver and defend persuasive technical presentations based on selected reliable evidence to the target audiences

• Demonstrate problem solving

• Assess and review progress

• Demonstrate independent learning research skills

• Apply written communication skills.

• Apply analytical skills

• Demonstrate autonomous task planning and implementation.

Module content

An introduction to sound modelling; key research and commercial applications in physical modelling; the mass/spring/damper model; discrete time modelling and finite difference numerical methods; the 1-D wave equation and the digital waveguide travelling wave solution; the plucked string digital waveguide model; speech modelling and the vocal tract; room models, geometric acoustics and reverb simulation; virtual analogue models; MATLAB implemented audio signal processing applications.

Automated segmentation of music and audio signals via onset detection; extraction of music related features using spectral analysis of audio such as chromagrams; melody analysis/metrics using pitch distributions and distance metrics such as cross correlation and Dynamic Time Warping; speech analysis of formants using MFCC; machine learning for classification and production of audio using K-Nearest Neighbours, Deep Neural Networks and Generative Adversarial Networks.

Specific module content will reflect most recent trends from research and industry but will typically include:

Physical modelling: exploring how acoustic systems can be analysed and modelled mathematically and implemented efficiently as discrete time equivalents to give new ways to synthesise and process sound.

Virtual analogue: demonstrating how legacy analogue audio devices can be modelled and implemented as digital equivalents.

Machine Learning: introducing how artificial intelligence, machine learning and data is influencing the development of next generation audio systems.

Music Information Retrieval (MIR): how to design and implement MIR systems, their real-world application and use as a tool to understand music-making and listening from different perspectives.

Voice and speech: human speech and modelling the vocal tract, the musical and acoustic features that emerge and their analysis.

Key examples from industry and research will be explored with guest lecture content bringing current insights from individuals working at the forefront of their respective fields.

Indicative assessment

Task	% of module mark
Essay/coursework	100.0

Special assessment rules

None

Additional assessment information

The assessment will be a portfolio of the following elements:

Online Report and Supporting Audio Examples: An online report into one or more aspects of the module (or related module) content, focused on a more general audience and using original or third party examples to highlight key aspects. This will also provide a platform for demonstrating the outcome of individual audio processing experimental and practical work via the use of sound and multimedia examples.

Music/Sound Analysis Implementation Report and Audio Examples: A report, adhering to IEEE journal guidelines, on the implementation and testing of a suitable music/sound analysis metric. This may be based on the development of a new analysis tool or the systematic assessment of the limitations of existing software. The report will be accompanied with sound files as appropriate.

Indicative reassessment

Task	% of module mark
Essay/coursework	100.0

Module feedback

Formative Feedback

• Regular Practical Exercises allow the application of the theory covered in lectures in a guided and structured way and students will receive verbal (online or in person) help and feedback on methods and results.

• Lab Demonstrators support the delivery of practical sessions and they will be able to give help and additional feedback on responses to the weekly practical exercises.

• Weekly sessions summarise the content from online lectures and practical exercises in the previous week, and anticipate the subjects to be covered in the week ahead. This allows questions to be raised around any issues with the module content, and verbal feedback to be given on progress to date.

• Online communication (via Email or other online means) to the Module Coordinator with Questions / Comments will be answered as soon as possible.

• Opportunities to speak about individual assignment plans are enabled through any of the regular timetabled sessions and via online communication. The final practical session of the term is focused around receiving feedback on assessment planning.

Summative Feedback

A standardised feedback sheet will be received for each assessment for the module, showing the final module mark, the marks breakdown against each of the learning objective assessment criteria, and how these have met based on a set of predefined grade descriptors that will be supplied with the assessment brief. In addition, personalised feedback will be given highlighting three clear areas of strength and three areas for improvement. The final module mark will also be made available via eVision.

Indicative reading

Gareth Loy, “Musimathics: The Mathematical Foundations of Music: 2”, MIT Press, 2011, ISBN-13: 978-0262516563

Udo Zölzer (Ed), “DAFX: Digital Audio Effects”, John Wiley & Sons, 2022, DOI: 10.1002/9781119991298

Curtis Roads, “The Computer Music Tutorial (Second Edition)”, MIT Press, 2023, ISBN: 9780262044912

Knees P. Music Similarity and Retrieval : An Introduction to Audio- and Web-Based Strategies . (Schedl M, ed.). Springer; 2016.

Mu¨ller M. Fundamentals of Music Processing [Electronic Resource] : Audio, Analysis, Algorithms, Applications . Springer; 2015.

Virtanen, T., Plumbley, M.D. and Ellis, D.P.W. (2018) Computational analysis of sound scenes and events. Cham, Switzerland: Springer Verlag.