Prof. David Marshall and Dr. Kirill Sidorov established the Computational Music research within the School a few years ago following their innovative paper on “Music Analysis As a Smallest Grammar Problem”, ISMIR 2014 (http://www.terasoft.com.tw/conf/ismir2014/). We can successfully analyse the high-level musical structure. Furthermore, we can edit this structure to produce new music similar to the original and wide range of other possible applications exists, including automatic summarization and simplification; estimation of musical complexity and similarity, and plagiarism detection.
We are interested in many Music related problems that involve computational analysis especially using Machine Learning, Deep Learning etc. These could involve audio, computer music representations, imagery (Sheet music) or video, or mixtures of any of these.
Some possible example projects include (but are not exclusive):
· Music Structure Analysis in Audio – This project is to apply the smallest grammar problem to Audio.
· Reverse Engineering Audio Synthesiser Sounds – The basic idea is given a set of sample recordings of a sound can you determine the appropriate parameters of a given synthesiser to best approximate the set of sounds.
· Sample based Musical Instrument Transcription, Annotation or Automatic DAW automation – The basic idea here is to use a large range of sample libraries to machine learn playing style and instrument type. Sample libraries can be programmed to produce a wide range of realistic instrument articulations allowing for large data samples that can be deployed in Machine Learning methods to allow for music transcription or annotation on a large scale. Programming Digital Audio Workstations (DAW) is highly work intensive. Articulations can be programmed into sample libraries to allow for the machine learning of DAW Automation.
· Deep Learning Guitar Tunings and Transcription – Guitars can be tuned to a a variety of tunings. This is especially common in certain genres of music, notably folk music and heavy metal, for example. The basic idea here is to take recordings on music played in known guitar tunings. I have database of known songs in know tunings by a variety of artists and have the associated recordings. Related to the previous point, sample libraries can also be used to transcribe guitar styles, techniques etc. to allow for complete guitar transcription.
· Dictionary Learning for Music Synthesis – By aligning a computer music representation of music with and audio rendition mappings can be learnt from such correlations o enable the synthesis on new audio from unseen new input audio. The styles of a particular player could be capture, enabling the resurrection of dead musicians to play new music: What would Jimi Hendrix sound like playing Metallica?
· Hearing the Future: Predicting the next piece of audio or musical data.
· Audio Behaviour Transfer – GANs (Generative adversarial networks) and other similar deep networks have proved popular in a variety of behavioural transfer and synthesis tasks. The basic idea to explore the use of such networks to transfer the characteristics of one audio file to another. Example applications include:
o Changing the human voice: accents, male <-> female
o Transferring the recording characteristics of one recording to another: E.g. room ambience, EQ.
o Changing the sound of an instrument. E.g. electric guitar to acoustic guitar
The above are merely suggestions of possible PhD projects and variations on the themes of the above and, indeed, any related are ideas. Please get in touch: [Email Address Removed] to discuss any such potential PhD projects.
Key words: Computer Music / Machine Learning / Artificial Intelligence / Digital Audio / Digital Signal Processing
Academic criteria: A 2:1 Honours undergraduate degree or a master's degree, in computing or a related subject. Applicants with appropriate professional experience are also considered. Degree-level mathematics (or equivalent) is required for research in some project areas.
Applicants for whom English is not their first language must demonstrate proficiency by obtaining an IELTS score of at least 6.5 overall, with a minimum of 6.0 in each skills component.
Application Information: If you would like to be considered for the School Funded Application, please submit your application before the 30th June 2021.
In the funding field of your application, insert “I am applying for 2021 PhD Scholarship in Computer Science and Informatics”, and specify the project title and supervisors of this project in the text box provided.
Apply online: https://www.cardiff.ac.uk/study/postgraduate/research/programmes/programme/computer-science-and-informatics - Please read the "How to apply" instructions carefully prior to application.