Artificial Neural Network Prediction Of Wavelet Sub-Bands For Audio Compression
Popular Tags
No tags have been applied to this document.
BibTeX | Add To MetaCart
author = {Robert Barry},
title = {Artificial Neural Network Prediction Of Wavelet Sub-Bands For Audio Compression},
year = {}
}
Bookmarks
OpenURL
Abstract:
A discrete waveform, such as an audio recording, can be represented by a series of wavelet sub-bands, produced by passing the waveform through a discrete wavelet transfer function. The audio recording chosen is considered to be easy listening music, and it is hoped that the wavelet sub-bands generated will possess an inherent predictability. Experimentation is done on the software level, using Matlab to calculate the wavelet sub-bands of the chosen waveform, and to train a backpropagation neural network using these wavelet sub-bands. This thesis investigates the possibility of predicting the N+1 th sub-band of a wavelet transform given the first N sub-bands, with the intention of adequately reconstructing the original waveform.
Citations
| 80 | Communication systems – Haykin - 2001 |
| 1 | Codebook image compression with neural networks”, B.Sc. Thesis – Dansereau - 1995 |
| 1 | Grondelle, “A Matlab implementation of neural networks – van - 1997 |
| 1 | to use the Matlab wavelet toolbox” (web page), Scientific Computing I – Petersdorff, “How - 1999 |
| 1 | Eric Weisstein’s world of mathematics” (web – Weisstein, Research, et al. - 1996 |
