Two voices : the marginalised body as a voice & Imperfect Offerings a novel

This thesis investigates the phenomenon of self-harm as a form of political protest using two different, but complementary, methods of inquiry: a theoretical research project and a novel. Through these two approaches, to the same research problem, I examine how we can re-position the body that self-harms in political protest from weapon to voice; and in doing so find a path towards ethical and equitable dialogue between marginalised and mainstream communities. The theoretical, or academic, portion of the thesis examines self-harm as protest, positing these acts as a form of tactical selfharm, and acknowledge its emergence as a voice for the otherwise silenced in the public sphere. Through the use of phenomenology and feminist theory I examine the body as site for political agency, the circumstances which surround the use of the body for protest, and the reaction to tactical self-harm by the individual and the state. Using Bakhtin’s concept of dialogism, and the dialogic space I propose that by ‘hearing’ the body engaged in tactical selfharm we come closer to entering into an ethical dialogue with the otherwise silenced in our communities (locally, nationally and globally). The novel, Imperfect Offerings, explores these ideas in a fictional world, and allows me to put faces, names and lives to those who are compelled to harm their bodies to be heard. Also using Bakhtin’s framework I encourage a dialogue between the critical and creative parts of the thesis, challenging the traditional paradigm of creative PhD projects as creative work and exegesis.

Rephrasing voice : art, practice-led research and the limits and sites of articulacy

While my PhD is practice-led research, it is my contention that such an inquiry cannot develop as long as it tries to emulate other models of research. I assert that practice-led research needs to account for an epistemological unknown or uncertainty central to the practice of art. By focusing on what I call the artist's 'voice,' I will show how this 'voice' is comprised of a dual motivation—'articulate' representation and 'inarticulate' affect—which do not even necessarily derive from the artist. Through an analysis of art-historical precedents, critical literature (the work of Jean-François Lyotard and Andrew Benjamin, the critical methods of philosophy, phenomenology and psychoanalysis) as well as of my own painting and digital arts practice, I aim to demonstrate how this unknown or uncertain aspect of artistic inquiry can be mapped. It is my contention that practice-led research needs to address and account for this dualistic 'voice' in order to more comprehensively articulate its unique contribution to research culture.

Lip detection for audio-visual speech recognition in-car environment

Acoustically, car cabins are extremely noisy and as a consequence audio-only, in-car voice recognition systems perform poorly. As the visual modality is immune to acoustic noise, using the visual lip information from the driver is seen as a viable strategy in circumventing this problem by using audio visual automatic speech recognition (AVASR). However, implementing AVASR requires a system being able to accurately locate and track the drivers face and lip area in real-time. In this paper we present such an approach using the Viola-Jones algorithm. Using the AVICAR [1] in-car database, we show that the Viola- Jones approach is a suitable method of locating and tracking the driver’s lips despite the visual variability of illumination and head pose for audio-visual speech recognition system.

Cascading appearance-based features for visual voice activity detection

The detection of voice activity is a challenging problem, especially when the level of acoustic noise is high. Most current approaches only utilise the audio signal, making them susceptible to acoustic noise. An obvious approach to overcome this is to use the visual modality. The current state-of-the-art visual feature extraction technique is one that uses a cascade of visual features (i.e. 2D-DCT, feature mean normalisation, interstep LDA). In this paper, we investigate the effectiveness of this technique for the task of visual voice activity detection (VAD), and analyse each stage of the cascade and quantify the relative improvement in performance gained by each successive stage. The experiments were conducted on the CUAVE database and our results highlight that the dynamics of the visual modality can be used to good effect to improve visual voice activity detection performance.

Shared leadership in Queensland schools : a collective case study

Shared leadership has been identified as a key governance base for the future of government and Catholic schools in Queensland, the state’s two largest providers of school education. Shared leadership values the contributions that many individuals can make through collaboration and teamwork. It claims to improve organisational performance and reduce the increasing pressures faced by principals. However despite these positive features, shared leadership is generally not well understood, not well accepted and not valued by those who practice or study leadership. A collective case study method was chosen, incorporating a series of semi-structured interviews with principals and the use of official school documents. The study has explored the current understanding and practice of shared leadership in four Queensland schools and investigated its potential for use.

On collective unintelligence

The idea of collective unintelligence is examined in this paper to highlight some of the conceptual and practical problems faced in modeling groups. Examples drawn from international crises and economics provide illustrative problems of collective failures to act in intelligent ways, despite the inputs and efforts of many skilled and intelligent parties. Choices made of “appropriate” perceptions, analysis and evaluations are examined along with how these might be combined. A simple vector representation illustrates some of the issues and creative possibilities in multi-party actions. Revealed as manifest (un-)intelligence are the resolutions of various problems and potentials that arise in dealing with the “each and all” of a group (wherein items are necessarily non-parallel and of unequal valency). Such issues challenge those seeking to model collective intelligence, but much may be learned.

The QUT-NOISE-TIMIT corpus for the evaluation of voice activity detection algorithms

The QUT-NOISE-TIMIT corpus consists of 600 hours of noisy speech sequences designed to enable a thorough evaluation of voice activity detection (VAD) algorithms across a wide variety of common background noise scenarios. In order to construct the final mixed-speech database, a collection of over 10 hours of background noise was conducted across 10 unique locations covering 5 common noise scenarios, to create the QUT-NOISE corpus. This background noise corpus was then mixed with speech events chosen from the TIMIT clean speech corpus over a wide variety of noise lengths, signal-to-noise ratios (SNRs) and active speech proportions to form the mixed-speech QUT-NOISE-TIMIT corpus. The evaluation of five baseline VAD systems on the QUT-NOISE-TIMIT corpus is conducted to validate the data and show that the variety of noise available will allow for better evaluation of VAD systems than existing approaches in the literature.

Fictionalising oral history : narrative analysis, voice and identity

Details of a project which fictionalises the oral history of the life of the author's polio-afflicted grandmother Beth Bevan and her experiences at a home for children with disabilities are presented. The speech and language patterns recognised in the first person narration are described, as also the sense of voice and identity communicated through the oral history.

The citizen's voice : Albert Hirschman's Exit, Voice and Loyalty and its contribution to media citizenship debates

This paper analyses Albert Hirschman's Exit, Voice and Loyalty (Hirschman 1970) as a basis for understanding the relationship between media and citizenship. It considers the significance of Hirschman's concept of voice in relation to media policy, media participation through user-created content, and the rise of 'citizen media' and 'citizen journalism'. It associates these developments with a 'de-centering' of both media practice and media studies, as considered by Couldry (2006a, 2006b). It concludes by suggesting that voice and participation, rather than citizenship, may constitute a more suitable foundation for understanding new digital media initiatives.

Models of collective cell spreading with variable cell aspect ratio : a motivation for degenerate diffusion models

Continuum diffusion models are often used to represent the collective motion of cell populations. Most previous studies have simply used linear diffusion to represent collective cell spreading, while others found that degenerate nonlinear diffusion provides a better match to experimental cell density profiles. In the cell modeling literature there is no guidance available with regard to which approach is more appropriate for representing the spreading of cell populations. Furthermore, there is no knowledge of particular experimental measurements that can be made to distinguish between situations where these two models are appropriate. Here we provide a link between individual-based and continuum models using a multi-scale approach in which we analyze the collective motion of a population of interacting agents in a generalized lattice-based exclusion process. For round agents that occupy a single lattice site, we find that the relevant continuum description of the system is a linear diffusion equation, whereas for elongated rod-shaped agents that occupy L adjacent lattice sites we find that the relevant continuum description is connected to the porous media equation (pme). The exponent in the nonlinear diffusivity function is related to the aspect ratio of the agents. Our work provides a physical connection between modeling collective cell spreading and the use of either the linear diffusion equation or the pme to represent cell density profiles. Results suggest that when using continuum models to represent cell population spreading, we should take care to account for variations in the cell aspect ratio because different aspect ratios lead to different continuum models.

Noise robust voice activity detection using features extracted from the time-domain autocorrelation function

This paper presents a method of voice activity detection (VAD) for high noise scenarios, using a noise robust voiced speech detection feature. The developed method is based on the fusion of two systems. The first system utilises the maximum peak of the normalised time-domain autocorrelation function (MaxPeak). The second zone system uses a novel combination of cross-correlation and zero-crossing rate of the normalised autocorrelation to approximate a measure of signal pitch and periodicity (CrossCorr) that is hypothesised to be noise robust. The score outputs by the two systems are then merged using weighted sum fusion to create the proposed autocorrelation zero-crossing rate (AZR) VAD. Accuracy of AZR was compared to state of the art and standardised VAD methods and was shown to outperform the best performing system with an average relative improvement of 24.8% in half-total error rate (HTER) on the QUT-NOISE-TIMIT database created using real recordings from high-noise environments.