Page 1 |
Save page Remove page | Previous | 1 of 150 | Next |
|
small (250x250 max)
medium (500x500 max)
large ( > 500x500)
Full Resolution
All (PDF)
|
This page
All
Subset |
STATISTICAL ENHANCEMENT METHODS FOR
IMMERSIVE AUDIO ENVIRONMENTS AND COMPRESSED AUDIO
by
Demetrios Cantzos
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(ELECTRICAL ENGINEERING)
December 2008
Copyright 2008 Demetrios Cantzos
Object Description
| Title | Statistical enhancement methods for immersive audio environments and compressed audio |
| Author | Cantzos, Demetrios |
| Author email | cantzos@usc.edu; cantzos@gmail.com |
| Degree | Doctor of Philosophy |
| Document type | Dissertation |
| Degree program | Electrical Engineering |
| School | Viterbi School of Engineering |
| Date defended/completed | 2008-10-20 |
| Date submitted | 2008 |
| Restricted until | Unrestricted |
| Date published | 2008-12-09 |
| Advisor (committee chair) | Kyriakakis, Chris |
| Advisor (committee member) |
Narayanan, Shrikanth Alexander, Kenneth |
| Abstract | Over the past years, audio enhancement methods have been the center of research focus as audio reproduction systems become ever more popular. Portable music players, internet radio and multichannel surround systems are all evidence of the rapid evolution of audio rendering algorithms and their applications. In this new environment, the primary role of audio enhancement algorithms is to improve on the performance and resource allocation of audio rendering systems given application-specific transmission bandwidth and data storage constraints.; The first key issue that this work attempts to address is the enhancement of multichannel audio signals. Inherently, multichannel audio systems incur high transmission and storage requirements which stem directly from the large number of audio rendering channels. We employ audio conversion methods to transform a given audio channel to another one at the cost of small transmission or storage overhead. In essence, a method is proposed on resynthesizing a large number of channels of any multichannel signal set from only one channel. Applications of this method can be found in transmission of multichannel audio over the Internet and storage of multichannel audio data.; The audio enhancement methods developed for multichannel scenarios are further extended on compressed stereophonic or monophonic audio signals. Low bitrate compression algorithms are extremely widespread since they allow for efficient storage and transmission of audio. Nevertheless, signal degradation is inevitable under low bitrate compression and thus a method on enhancing the quality of the degraded signal is presented. The proposed algorithm attempts to improve the quality of a compressed, degraded audio signal by means of a statistical, subband-adaptive transformation derived between the degraded and desired signals. The primary design consideration of such an algorithm is the use of minimum amount of parameters to represent the transformation as well as scalability to accommodate for variable enhancement performance. An extension of this algorithm that enables it to operate in a bandwidth extension mode is also presented. In that case the quality improvement is restricted to the higher end of the audio spectrum, while the lower frequencies are handled by a standard compression scheme.; A rather more challenging task for audio enhancement schemes is related to audio restoration. In that context, audio enhancement is concerned with improving the quality of audio signals assuming that any prior information of the desired audio signal has been lost. The missing information is acquired by using corpus-based techniques related to large training data of similar music genres. The purpose of such a scheme is to find the most suitable transformation among a large number of available transformations that will convert a degraded signal to a signal of better audio quality. The amount of parameters used to realize these transformations is irrelevant for this scenario but the selection of the appropriate transformation is crucial. |
| Keyword | audio enhancement; audio synthesis; statistical transformation |
| Language | English |
| Part of collection | University of Southern California dissertations and theses |
| Publisher (of the original version) | University of Southern California |
| Place of publication (of the original version) | Los Angeles, California |
| Publisher (of the digital version) | University of Southern California. Libraries |
| Provenance | Electronically uploaded by the author |
| Type | texts |
| Legacy record ID | usctheses-m1898 |
| Rights | Cantzos, Demetrios |
| Repository name | Libraries, University of Southern California |
| Repository address | Los Angeles, California |
| Repository email | http://www.usc.edu/isd/libraries/services/ask_a_librarian/email/ |
| Filename | etd-Cantzos-2514 |
| Archival file | uscthesesreloadpub_Volume48/etd-Cantzos-2514.pdf |
Description
| Title | Page 1 |
| Full text | STATISTICAL ENHANCEMENT METHODS FOR IMMERSIVE AUDIO ENVIRONMENTS AND COMPRESSED AUDIO by Demetrios Cantzos A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (ELECTRICAL ENGINEERING) December 2008 Copyright 2008 Demetrios Cantzos |
Comments
Post a Comment for Page 1
