Date of Award

2010

Publication Type

Doctoral Thesis

Degree Name

Ph.D.

Department

Mechanical, Automotive, and Materials Engineering

Keywords

Engineering, Mechanical.

Supervisor

Novak, Colin (Mechanical, Automotive and Materials Engineering)

Rights

info:eu-repo/semantics/openAccess

Abstract

Experimental studies have shown that for short gaps between 2 to 5 ms, the perceived loudness is higher than for uninterrupted noise presented to the ear. Other studies have also shown that the present temporal integration models for the calculation of time varying loudness do not adequately account for short duration phenomena. It has been proposed that the multiple look approach is a more applicable method for describing these short term circumstances. This approach breaks a sound into small durations or looks having length of 1 ms which allows for the intelligent processing of the looks and decision making depending on the nature of the stimulus. However, present technologies (i.e. FFT) are not adequate to deal with short duration sounds across the entire frequency spectra. A compromised approach is taken here to account for perceived loudness levels for sounds in the presence of gaps while using an integration model. This approach is referred to as a multiple look gap adjustment model. A model and software code was developed to take a recorded sound presented to the ear and process it into individual looks which are then examined for the presence of gaps ranging in length between 1 to 10 ms. If gaps are found, an appropriate gap adjustment is applied to the sound. The modified stimulus is subsequently evaluated for loudness level using a model which relies on temporal integration. The multiple look model was tested using several sounds including mechanical and speech sounds and was found to perform as intended. While recommendations for improvement and further study are included, the application of the model has shown particular merit for perceptional analysis of sounds involving speech.

Share

COinS