Use this URL to cite or link to this record in EThOS:
Title: Size discrimination of transient signals
Author: O'Meara, N.
ISNI:       0000 0004 2731 8430
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 2012
Availability of Full Text:
Access from EThOS:
Access from Institution:
The importance of spectral cues in size discrimination of transient signals was investigated, and a model for this ability, tAIM, was created based on the biological principles of human hearing. A psychophysics experiment involving 40 participants found that the most important cue for size discrimination of transient signals, created by striking different sizes of polystyrene spheres, was similar to that of speakers listening to vowels – the relative positions of the resonances between comparison signals. It was found possible to scale the sphere signals in order to confuse listeners into believing the signal source was a different size, but two methods of scaling signals in order to sound the same size as another proved inconclusive, suggesting the possibility that transient signals cannot be scaled in a linear fashion as has been shown possible for vowels. Filtering the signals in a number of different ways found that the most important cue in size discrimination of transient signals is the difference between the most prominent resonances available in the spectra of the comparison signals. A model of the auditory system using the dynamic compressive Gammachirp filterbank, and based on the well-known AIM, was created to produce auditory images of transient signals that could be normalised for size. Transient-AIM, or tAIM used the Mellin transform to produce images that showed size normalisation was possible due to the spectral envelope similarities across the sizes of the spheres. tAIM was extended to carry out size discrimination of the spheres using the information contained within the Mellin images. There was a systematic association between Mellin phase and size of objects of various shapes, which suggests that tAIM is able to infer object size from sound recordings of objects being struck.
Supervisor: Bleeck, Stefan Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: TA Engineering (General). Civil engineering (General)