This thesis presents an investigation into developing musical systems with an Unconventional Computing substrate. Computer musicians have found it difficult to access the field of Unconventional Computing, which is likely due to its resource-intensive and complex nature. However, ongoing research is establishing the myxomycete Physarum polycephalum as a universally-accessible and versatile biological computing substrate. As such, the organism is a potential gateway for computer musicians to begin experimenting with aspects of Unconventional Computing. Physarum polycephalum, in its vegetative plasmodium form, is an amorphous unicellular organism that can respond with natural parallelism to the environmental conditions that surround it. This thesis explores the challenges and opportunities related to developing musical systems with Physarum polycephalum. As this area of inquiry is in its infancy, the research took inspiration from a common approach in Unconventional Computing: a journey of exploration and discovery. This journey consisted of a selection of waypoints that provided direction while allowing the research to explore applications of Physarum polycephalum in order to establish how it may be useful in Computer Music. These waypoints guided the research from adapting established prototypes for musical application to developing purpose-made musical demonstrators for use outside of the laboratory. Thus, the thesis reports on a series of Computer Music systems that explore one or more features of Physarum polycephalum's behaviour and physiology. First, the text presents an approach to algorithmic composition that exploits the organism's ability to form and reconfigure graph-like structures. Next, the thesis reports on systems that harness the plasmodium's electrical potential oscillations for sound synthesis and compositional tools. Finally, the thesis presents musical devices that encompass living plasmodium as electrical components. Where applicable, the thesis includes artefacts from demonstrations of these systems, some of which were developed in collaboration with a composer. The findings from this journey demonstrate that Physarum polycephalum is an appropriate substrate for computer musicians wanting to explore Unconventional Computing approaches creatively. Although Physarum polycephalum is relatively robust as a biological substrate, several obstacles arose during this project. This research addressed such obstacles by reviewing and selecting approaches that maintained the organism's accessibility to computer musicians. As a result, the work suggests methods for developing systems with the organism that are practical for the average music technologist and also beneficial to the wider group of scientists investigating Physarum polycephalum for other purposes.

This thesis traces the history and development of the facilities for electronic music in the UK. It covers the early attempts to experiment with electronic music and create studios in less than ideal circumstances and the subsequent bids to create a national centre. It also covers some elements of worldwide development of electronic music and sound recording, in particular those which occurred before 1965. The thesis calls upon non-traditional sources and the author was able to access many documents in the personal archives of electronic music pioneers. There is substantial reference to committees and societies for electronic music and their effects on the development of facilities for electronic music in the UK. Some of the early pioneers are studied in detail; these include Daphne Oram, Tristram Cary and Hugh Davies. Unprecedented access to information on Hugh Davies and Daphne Oram was provided by the family estates of these recently deceased composers. This allowed the author to gain valuable insight into the working patterns and methodology of these composers. Many references to later pioneers such as Trevor Wishart are also made but the focus remains on the facilities available to composers rather than the composers and their works.

This thesis examines the practices and techniques involved with particular electronic instruments and proposes an archaeological approach to reconsider the ways in which noise can communicate various details of instrument design and practice to the listener. I present two case studies concerning electronic music practice using repurposed devices - stepped filters - and by combining a detailed material analysis of the instruments with interviews, video and other evidence, I document the practices involved with their use. By rebuilding these instruments, and designing and building other devices, I test my hypotheses through my own practice, and by doing so I refine my results and extend my composition, performance practice and technical design skills to include valuable lessons learned through this research. The portfolio engages with the three archaeological levels (Listening Situation, Reproduction Stage, Production Environment) and the three areas of the production continuum (Composition, Performance Practice, Technical Design) and through sound installations, crafted media, recorded performances, and the documentation of devices designed for these pieces, it supports the thesis through experimentation and incorporation of results through reflective practice.

The size of easily-accessible libraries of digital music recordings is growing every day, and people need new and more intuitive ways of managing them, searching through them and discovering new music. Musical emotion is a method of classification that people use without thinking and it therefore could be used for enriching music libraries to make them more user-friendly, evaluating new pieces or even for discovering meaningful features for automatic composition. The field of Emotion in Music is not new: there has been a lot of work done in musicology, psychology, and other fields. However, automatic emotion prediction in music is still at its infancy and often lacks that transfer of knowledge from the other fields surrounding it. This dissertation explores automatic continuous dimensional emotion prediction in music and shows how various findings from other areas of Emotion and Music and Affective Computing can be translated and used for this task. There are four main contributions. Firstly, I describe a study that I conducted which focused on evaluation metrics used to present the results of continuous emotion prediction. So far, the field lacks consensus on which metrics to use, making the comparison of different approaches near impossible. In this study, I investigated people’s intuitively preferred evaluation metric, and, on the basis of the results, suggested some guidelines for the analysis of the results of continuous emotion recognition algorithms. I discovered that root-mean-squared error (RMSE) is significantly preferable to the other metrics explored for the one dimensional case, and it has similar preference ratings to correlation coefficient in the two dimensional case. Secondly, I investigated how various findings from the field of Emotion in Music can be used when building feature vectors for machine learning solutions to the problem. I suggest some novel feature vector representation techniques, testing them on several datasets and several machine learning models, showing the advantage they can bring. Some of the suggested feature representations can reduce RMSE by up to 19% when compared to the standard feature representation, and up to 10-fold improvement for non-squared correlation coefficient. Thirdly, I describe Continuous Conditional Random Fields and Continuous Conditional Neural Fields (CCNF) and introduce their use for the problem of continuous dimensional emotion recognition in music, comparing them with Support Vector Regression. These two models incorporate some of the temporal information that the standard bag-of-frames approaches lack, and are therefore capable of improving the results. CCNF can reduce RMSE by up to 20% when compared to Support Vector Regression, and can increase squared correlation for the valence axis by up to 40%. Finally, I describe a novel multi-modal approach to continuous dimensional music emotion recognition. The field so far has focused solely on acoustic analysis of songs, while in this dissertation I show how the separation of vocals and music and the analysis of lyrics can be used to improve the performance of such systems. The separation of music and vocals can improve the results by up to 10% with a stronger impact on arousal, when compared to a system that uses only acoustic analysis of the whole signal, and the addition of the analysis of lyrics can provide a similar improvement to the results of the valence model.

The basis of this research project stems from reflections upon the process of composition for electroacoustic audio-visual music. These are fixed media works in which sound and image materials are accessed, generated, explored and configured in creation of a musically informed audio-visual expression. Within the process of composition, the composer must decide how to effectively draw relationships between these time based media and their various abstract and mimetic materials. This process usually has no codified laws or structures and results in relationships that are singular to the individual artworks. The composer uses their own experience and intuition in assessing how best to associate sounds and images and they will use their own interpretation of the materials to evaluate the how successful they are in realising their intentions. But what is there to say that the interpretation made by the composer bares any resemblance to interpretations made by audiences? The current research sought to assess any trends or commonalities in how people interpret such works. Utilising a combination of empirical research, composition and scholarly study, the project investigated various theoretical approaches to interpretation and the occurrence of correlation between compositional intention and audience interpretation. Models from different theoretical disciplines were combined in order to build up a picture of the processes involved in making interpretations, and to aid in the rationalisation of empirical data. The application of three methodological approaches allowed for the topic to be considered from a diversity of perspectives, and for triangulation to take place in confirmation of the research outcomes. The way in which individuals build up interpretations from non-codified abstract and mimetic materials also provided a suitable case study for the critique and assessment of various theoretical approaches to interpretation. The project challenges structuralist approaches to interpretation, drawing together theoretical materials and empirical research findings in support of a post-structrualist model of interpretation that demonstrates the absolutely vital role played by context - the framing of the artwork in the consciousness of the individual audience member.

There is an arguable lack of activity and interest in the analysis of electroacoustic music when compared to its composition and performance. The absence of a strong and active analytical community is very concerning, as it should be a fundamental part of any larger musical community that wishes for works to be performed and discussed in later years. The problems that face electroacoustic music analysis are that there is no consensus or single analytical tool/methodology that dictates how such an activity should be undertaken. Rather than attempting to appropriate existing tools meant for traditional musics or create a new universal one this thesis will argue that a new culture should be adopted that promotes different opinions on the subject of electroacoustic music analysis, as opposed to defining a consensus as to how it should be conducted. To achieve this the thesis will: evaluate and critique what constitutes and defines electroacoustic music analysis; provide a general and flexible procedure to conduct an analysis of an electroacoustic work; develop a set of criteria and terms to cross-examine the current analytical tools for electroacoustic music in order to define the gaps in the field and to identify pertinent elements within electroacoustic works; analyse a number of electroacoustic works to test and implement the ideas raised within this thesis; and finally the concept of an analytical community (in which such a culture could exist) is outlined and implemented with the creation of the OREMA (Online Repository for Electroacoustic Music Analysis) project. This universal approach will cover both epistemological and ontological levels of electroacoustic music analysis. All of the concepts raised above are interlinked and follow the main hypothesis of this thesis: • There is no one single analysis that can fully investigate a work; • Analyses are a perspective on a work, ultimately formed through the subjective perception of the analyst; • These perspectives should be shared with other practitioners to help develop a better understanding of the art form. This PhD study was part of the New Multimedia Tools for Electroacoustic Music Analysis project (2010-2013) funded by the Arts and Humanities Research Council (UK). Other outcomes of that project included the various analysis symposiums held at De Montfort University in Leicester and the electroacoustic analysis software EAnalysis created by Pierre Couprie.

L’étude de la musique mixte est caractérisée par un dualisme d’approches, ceux qui étudient les dispositifs et ceux qui utilisent les stratégies analytiques perceptives issues de la musique électroacoustique. Pour surmonter ce dualisme, cette thèse esquisse une typologie à partir de la perception de l’interaction des composantes instrumentale et électronique : les œuvres mixtes sont considérées comme des «œuvres hybrides» juxtapositionnelles, synthétiques et transformationnelles. Afin de conjuguer les aspects de la perception et de la prescription dans la musique mixte, les notions de geste et texture s’avèrent pertinentes. Une approche analytique holistique est proposée. La définition d’un objet spécifique, le « son mixte », nous permet d’interpréter d’une manière originale la spectromorphologie de Denis Smalley et d’examiner ses fondements épistémologiques. Les critères perceptifs de la typologie s’avèrent fondés sur une répartition fonctionnelle des débuts, entretiens et extinctions sonores des « sons mixtes » entre la partie instrumentale et électronique. Cette perspective est validée par l’analyse comparative de cinq cas d’étude : Sopiana et Aulodie de François-Bernard Mâche, Pluton de Philippe Manoury, Traiettoria de Marco Stroppa et Mixtur de Karlheinz Stockhausen. / The study of mixed music is characterized by a dualist approach: on the one hand, an analysis that merely focuses on technical meansand, on the other hand, research resorting to perceptive strategies in the analysis of electroacoustic music. In order to overcome thistraditional dualism, this work sketches a perceptive typology of the interaction between instrumental and electronic components, mixedmusic works are considered juxtapositionals, synthetics and transformationals. In doing so, the thesis advances an analytical holisticapproach in order to combine the concepts of perception and prescription in mixed music, where the notions of gesture and texture areparticularly relevant. The definition of a specific objet, i.e. the “mixed sound”, allows for an original interpretation of Denis Smalley’sspectromorphology and its epistemological framework. As a result, the typological perceptive criteria here proposed are based on thefunctional distribution of the attacks, sustains and decays of “mixed sounds” between the instrumental and the electronic parts. Thishypothesis is illustrated through the detailed analysis of five case studies: François-Bernard Mâche’s Sopiana and Aulodie, PhilippeManoury’s Pluton, Marco Stroppa’s Traiettoria and Karlheinz Stockhausen’s Mixtur.