No description available.

No description available.

No description available.

No description available.

This thesis describes the background, motivation, development and testing of a Spatial Audio Attribute Listener Training System (SAALTS). Previous studies in spatial audio attributes are examined, and Rumsey's Scene-Based Paradigm is found to be the only one that provides a rigorous approach for describing spatial audio scenes. Issues that would cause complications when implemented in a training system are resolved with the development of the Simplified Scene-Based Paradigm (SSBP) which can be used in the description of a wide range of spatial audio scenes for normative, product evaluation or training investigations. A pilot study to ascertain the effectiveness of a spatial audio attribute training system based upon the training of ranking tasks, and its transferral to tasks involving the rating of spatial audio attributes is reported. As a result of the pilot study, it is concluded that training naive listeners in the concept and judgement of spatial audio attributes as outlined in the SSBP is possible. This training, however is only found to have transferred as an increase in the range of the scale used by a (potentially more motivated) sub-set of the trained listeners. Informed by the pilot study and the literature on transfer and motivation, the Spatial Audio Attribute Listener Training System (SAALTS) detailed in this thesis employs the following elements: a tutorial explaining the SSBP and its importance; active learning using the Spatial Audio Attribute Toolkit (SAAT) and self-guided training drills involving motivation-inspiring elements. As a result of the research described in the transfer investigation, it can now be concluded that training in spatial listening can improve performance in spatial audio evaluation tasks that transfers to similar tasks with similar and different stimuli. It was also found that the performance of SAALTS was comparable to a repetitive practice regime for the target task.

This thesis describes the development of a psychoacoustically optimised bandwidth limitation method for multichannel audio. The proposed method takes advantages from the hierarchical representation of multichannel audio signals. Several hierarchical transform techniques were reviewed and their applicability to be used in the proposed hierarchical bandwidth limitation method were verified using psychoacoustical studies. The Karhunen-Loeve Transform (KLT) outperformed other hierarchical transform techniques due to its ability of rearranging the information in a perceptually hierarchical manner. Another advantage of KLT is that the transform matrices are signal-content-dependent and are adaptive to the actual spatial characteristics of audio material. The perceptual effects of the adaptive KLT-based hierarchical bandwidth limitation were studied. The results showed that applying KLT to short-term multichannel audio signals and updating the KLT transform matrices adaptively over time will result in better audio quality compared with non-adaptive KLT for those multichannel audio programmes that had significantly varying statistical characteristics over time. In addition, the bandwidth allocation strategy in KLT-based hierarchical bandwidth limitation was optimised for 3/2 stereo format for two levels of overall bandwidth: 40 and 60 kHz. An evaluation has been made of the optimised KLT-based hierarchical bandwidth limitation algorithm by comparing it with other bandwidth limitation algorithms. The results showed that the KLT-based hierarchical bandwidth limitation with the optimal bandwidth allocation strategy provided higher audio quality than the other traditional bandwidth limitation algorithms used for the limitation of bandwidth of multichannel sound.

The purpose of this study was to map, for a 5-loudspeaker surround system, perceptual dimensions of preference and of audible differences between stimuli, using stimuli that have been created to vary physical aspects of signals that are theorized to influence localization-quality and spaciousness. 'Localization Quality' was defined as the value of re; the vector length in an equation describing summing localization. Spaciousness is varied by manipulating the playback material to change the arrival direction of the late (post-80 ms.) portion of the reverberation. In a scaled pair-comparison experiment subjects were asked to scale the magnitude of the audible difference between each pair of stimuli, and to judge which was preferred. 7 stimuli were presented in every possible pair, each stimulus was created to have one of three values of the localization parameter and one of three categories of arrival direction for the reverberation. The author has contributed to the understanding of the perceptual structure of sound quality for reproduced sound. Descriptors of auditory attributes for natural hearing were elicited, and these were useful for the evaluation of multi-channel surround sound audio. Seven rating scales were created which can be used to distinguish between stimuli, thus providing an indication of the ways stimuli are perceived to be different. Multi-dimensional Scaling (MDS) was applied to listener ratings of difference between stimuli and showed that the stimuli were discriminable on two perceptual dimensions. Multi-dimensional Scaling was used to perform a weighted unfolding analysis of listener ratings of preference, and this created a map that reveals the relative preference for each of the stimuli for each of the subjects, the preference for each of the stimuli for the group as a whole, and how subjects differ in the weighting each applies to the revealed preference-map dimensions.

No description available.

Spatial impression in both concert halls and reproduced sound has been identified as an important attribute of the listening experience. In this study, the synthesis and objective measurement of spatial impression in reproduced sound is examined. A novel, multichannel spatializing technique for musical synthesis has been developed that entailed the separation of the individual harmonics of a musical note that were spatially distributed over multichannel surround systems. Subjective testing of the techniques revealed that the perceived degree of spatial impression significantly increased as the angular spread of harmonics increased, however, extending the spatial spread beyond 90° did not significantly increase the perception of spatial impression. The concert hall measure of spatial impression, the interaural cross correlation coefficient (IACC) was used to objectively measure the effects of the spatializing techniques. The IACC measurements displayed a strong correlation to the subjective results. Further examination of the IACC measurement indicated the possibility of it’s adaptation to multichannel surround sound in general. A method of adapting IACC to reproduced sound was further developed that involved comparing IACC measurements taken in a concert hall to IACC measurements taken in reproduced versions of the same concert hall. The method was first conducted as a simulation using basic auralisation techniques. Real concert hall measurements and reproduction systems were then employed. Results showed that the method was able to discriminate between the spatial capabilities of a number of different surround sound systems and rank them in a predictable order. The results were further validated by means of a subjective test. In an attempt to sensitise the IACC measurement, the frequency dependency of IACC was investigated by means of a subjective test. The results indicated that a perceptually more accurate indication of spatial impression may be gained by applying a frequency-dependent weighting to IACC measurements. This may be useful in the spatial measurement of both reproduced sound and concert halls.