Generally, the thesis reports on a series of eight experiments investigating how the face is communicating laughter, when using spontaneous and dynamic laughter clips. More specifically, the experiments are investigating how the laughing face may aid the ability to hear the laughter sound: an audiovisual laughter effect. The basic methodology was borrowed from the area of audiovisual speech perception, a well established area of research investigating how the speaking face aids the ability to hear the spoken word. Evidence for audiovisual laughter perception was found in each of the eight experiments. However, another question was how the face is aiding the ability to hear the laughter sound. Experiments one and two simultaneously investigated into the possibility of configurational processing in audiovisual laughter. Some evidence was found to this effect in experiment two but not experiment one. Experiments three, four and five simultaneously investigated into the possibility of featural processing in audiovisual laughter. Evidence was found to this effect in each experiment. Experiments six, seven and eight simultaneously investigated into the possibility that dynamic information, rather than static information is intrinsic to the audiovisual laughter effect. Evidence in favour of this possibility was found in each experiment. Overall, the data would appear to suggest the moving mouth is the single most salient feature for audiovisual laughter, with particular importance given to the moving featural detail of the intra-oral region. However, if this information is missing, the remaining parts of the moving laughing face are still able to aid the hearing of the laughter sound. Future research is necessary to further highlight the processes of the new perceptual phenomenon of audio visual laughter.

The phenomenon of attentional capture by a unique yet irrelevant "singleton" distractor has typically been studied in visual search. In this thesis I ask whether attention can also be captured by auditory singletons in tasks of auditory search. Participants searched sequences of sounds for targets defined by frequency, intensity or duration. The presence of a "singleton" distractor that was unique on an irrelevant dimension (e.g. a low frequency singleton in search for a target of high intensity) was associated with search costs in both detection and discrimination tasks. However if the singleton feature coincided with the target item, search was facilitated. These results demonstrate attentional capture in the auditory domain. Further experiments showed that, like visual attentional capture, auditory attentional capture depends on participants adopting an "odd-one-out" strategy for attentional allocation. In addition, a final set of experiments confirmed that singleton interference could be found in tasks of visual sequential search, providing a direct visual analogue for the present auditory effects. Overall, this thesis establishes the phenomenon of auditory attentional capture, adding to a growing body of research demonstrating similarities between attentional processes in vision and hearing.

No description available.

No description available.

This thesis is concerned with cognitively-motivated representations of musical structure. Three problems are addressed, each related in terms of their focus on music as an object of perception, and in the application of geometrical methods of knowledge representation. The problem of pattern discovery in discrete representations of polyphonic music is first considered, and a heuristic proposed which seeks to assist musicological analysis by identifying patterns that may be salient in perception, from a large number of potential patterns. This work is based on geometric principles that are far removed from plausible psychological models of pattern induction, but the method is motivated by psychological evidence for the importance of invariance and repetition in perception. The second and third problems explicitly adopt a cognitive theory of representation, namely the conceptual space framework developed by Gärdenfors (2000). Within this framework, concepts can be represented geometrically within perceptually grounded quality dimensions, and where distance in the space corresponds to similarity. The second problem concerns the prediction of melodic similarity, and the theory of conceptual spaces is investigated in the novel context of point set representations of melodic structure, employing the Earth Mover's Distance metric (Rubner 2000). This work builds on the work of Typke (2007) concerning the application of Earth Mover's Distance to melodic similarity. Evaluation is performed with respect to published psychological data (Müllensiefen 2004), and the MIREX 2005 symbolic melodic similarity evaluation. The third problem concerns the conceptual representation of metrical structure, informed by the psychological theory of metre developed by London (2004). A symbolic formalisation of this theory is developed, alongside two geometrical models of metrical-rhythmic structure, which are evaluated within a genre classification task.

Echolocation involves obtaining information on an object by analysis of echoes reflected from it when an outgoing emission is directed towards it. This study investigates human ability to determine the left-right position of an object using echolocation in a specific set of conditions. Previous literature demonstrated that echolocation can be used to identify the spatial position of an object, though the acoustic cues used to achieve such a task have not been identified. Understanding whether subjects are able to use binaural cues for such a task is important because, if they can, then the determination of spatial position via echolocation can be generalised to real-life conditions. Furthermore, ISVR pilot studies have demonstrated that subjects may have had access to additional echolocation cues in previous studies. This study investigates the ability of subjects with minimal (procedural) training to determine the left-right position of an object in 'virtual echolocation' experiments in which cues unrelated to echoes from the emission were eliminated. Impulse response measurements taken when an emission was directed at a board positioned at 17· to the ears of an acoustic mannequin, KEMAR, were analysed and combined (convolved) with emission signals. Convolved stimuli were presented to subjects via insert earphones in three psychoacoustic experiments conducted by the author. Experiment 1 used 18 subjects to examine sensitivity at 0.9 m using broad band stimuli of 10-400 ms, showing that sensitivity increased with signal duration. By windowing out the start of each signal, it was also shown that the precedence effect, which generally reduces a subjects' sensitivity to binaural cues within echoes, does not affect ability at the distance measured. The experiment also demonstrated that small changes to board position could have dramatic effects on information useable within the echo, explaining why many subjects have been observed moving their heads when echolocating, a technique known as 'head scanning'. A set of 13 subjects were common to experiments 2 and 3, with participants using 400 ms-duration stimuli to identify board position. Experiment 2 showed sensitivity above chance levels for echolocater to object distances of 0.6 to 1.2 m using broadband stimuli. Experiment 3 applied diotic presentation and level-roving to examine whether subjects relied on non- binaural cues at 0.9 m. Frequency-filtering isolated any binaural cues used. Results indicated that high-frequency interaural level difference (lLD) was the main cue, though 2 trained musicians showed some ability to use non-binaural cues. It is concluded that subjects with minimal training can use echolocation to determine the left-right spatial position of a large reflective object placed at a 17 degree angle from them based primarily on high-frequency lLD, at least at a distance of 0.9 m. Thus, it should be possible to identify spatial position of an object outside of the laboratory via echolocation. Furthermore, increased sensitivity with duration indicates that subjects should try to maximize energy in their emission signal when echolocating. Further work should be conducted to isolate the acoustic cues used in other echolocation tasks and the virtual technique piloted here provides a mechanism for doing so.

Elderly listeners generally have more difficulty than young listeners understanding one talker in the presence of another. Some of this difficulty may be attributable to "informational masking", which is the extra listening difficulty due to competing speech in comparison to acoustically equivalent noise maskers. A range of methods are used to measure informational masking for young and elderly listeners.

Binaural masking release is a psychophysical phenomenon whereby the binaural properties of a sound signal (ie., the relationship of the sound reaching the two ears) can alleviate the masking effect of background noise. The largest release from masking occurs when the waveform of either the signal or the masker is inverted at one ear. The signal detection threshold for this so-called "antiphasic" condition can be as much as 12- 15 dB lower than the threshold for the condition in which both the signal and the masker are identical at the two ears (referred to as "homophasic" condition). The difference between the homophasic and antiphasic thresholds is known as the binaural masking level difference (BMLD). The aim of this thesis is to explore the neural mechanisms of binaural masking release in both humans and other mammals (guinea pigs). In the first experiment, electroencephalographic (EEG) techniques were used to develop an accurate and objective neural correlate of binaural masking release. A linear relationship was observed between EEG response size and signal level for stimuli presented in the homophasic and antiphasic configurations. Backward extrapolation of a linear regression function to zero response size gave a remarkably accurate estimate of the psychophysical detection thresholds. This held true not only for the average data, but also on an individual basis for most participants. This measure was then used to test the underlying assumption of models of binaural masking release. The Equalisation-Cancellation (EC) model is the dominant psychophysical model of this effect and assumes that the underlying processing is mediated by the responses of neurons tuned to the binaural properties of the noise. An 2 interaural time difference of ± 500 liS was imposed on the entire stimulus so that the signal was perceived as clearly lateralised towards one hemifield or the other. The results showed that the hemispheric distribution of the neural responses reflected the perceived lateralisation of the signal (Le., greater activity in the hemisphere contralateral to the signal's lateralisation). In contrast to the EC model, these data suggest the global neural response to binaural masking release conditions is dominated by the activity of neurons sensitive to the signal properties. The final experiment was designed investigate whether the neuronal responses measured globally in the EEG experiments reflect the coincidence-detecting behaviour observed in binaurally sensitive neurons in the midbrain. Single unit responses were measured in the primary auditory cortex of urethane anaesthetised guinea pigs to homophasic and antiphasic binaural masking release conditions. Individual auditory cortex neurons could signal the presence of the tone in the masking noise by either an increase or a decrease in discharge rate, in a manner generally consistent with their interaural time difference sensitivity. Across the sample of auditory cortex neurons the magnitude of the BMLD was of the order of that shown psychophysically in humans. A simple cross-correlation function was capable of explaining much of the data collected here. This physiological data suggests the coincidence-detecting behaviour of midbrain neurons is faithfully projected to the cortex. The EEG data showed that neurons sensitive to the ITD of the signal are recruited in binaural masking release processing. The combination of these results is not compatible with the current assumptions of the EC model. 3

No description available.

No description available.