No description available.

No description available.

M.Sc., Faculty of Science, University of the Witwatersrand, 2011 / Studies have shown that text-to-speech systems need detailed prosodic models of a language in order to ideally sound natural to native speakers of the language. A text-to-speech system developed for Sesotho needs to have tone implemented in it since Sesotho is a tonal language which uses pitch variations to distinguish lexical and/or grammatical meaning. In order to implement tone for a language such as Sesotho, it is necessary for a tone modeling algorithm to receive as input the tone labels of the syllables of a word. This allows the algorithm to predict the appropriate intonation of the word. The aim of our study is to improve a basic tone labeling algorithm that predicts tone labels using three Sesotho tonal rules. The application of this algorithm is restricted to polysyllabic verb stems. The research study involves implementing an extended tone labeling algorithm that implements four additional Sesotho tonal rules and extends its application to all the other parts of speech. The results of our study show that the extended tone labeling algorithm significantly improves the basic algorithm by increasing the number of matched tone labels. Furthermore, our study provides the basic step to tone modeling for languages such as Sesotho which do not mark tone labels in orthography.

"...music is an art that exists in point of time." Aaron Copland, What to Listen for in Music / Few definitions are adequate to describe music, but a "point of time" is a concept with which people are familiar. When musicians give concerts they try to create these points in a context, which allows the audience to observe each moment by itself. Concert practice has developed to define the edges of musical points, guided by cues such as clapping, pauses, and concert program notes. / This masters thesis investigates how to analyze concert recordings of Western music and their program notes to produce segments which best fit the boundaries of musical points. Modern segmentation techniques are reviewed and a new method specific to concert recordings is examined.

The process of remastering old recordings comprises two basic stages: the transfer of the material from the source to a modern format, and the subsequent restoration of the transferred material through various forms of signal processing. The transfer stage in particular requires an understanding of issues which are becoming increasingly less familiar to engineers as the science of recording progresses further into the digital era. To a lesser extent, the restoration stage involves the use of certain techniques and forms of processing which are specific to this application. / This thesis is intended as a reference for those recording engineers who occasionally undertake remastering projects, but who are not thoroughly acquainted with the many different situations and problems which can be involved. Emphasis is given to those areas which are likely to be least familiar to most engineers. / In order to enable the engineer to properly assess a given source, the evolution and characteristics of each type of source--cylinder, disc and tape--is surveyed. This is followed in each case by an examination of the preparation, equipment and method used in transferring the source. Finally, the various types of processing which can be applied to the transferred material are presented, focusing on the techniques and forms of signal processing which are specific to audio restoration.

This thesis attempts to chart the development of a Jamaican musical form known as dub. This development is considered primarily in terms of the island's encounter with a series of new playback, amplification, recording, and sound treatment technologies. Section I focuses on the formation of the Jamaican sound system (a network of powerful mobile discos) and its pivotal role in the birth of a fertile domestic record industry. Section II extends the investigation to the Jamaican recording studio and record industry. What distinguishes this work from others on Jamaican dub is its emphasis on technology, and theories of technology, within a geo-political framework. In Section I, this emphasis is most notably informed by the work of Harold Innis, Karl Marx and Lewis Mumford, with Marshall McLuhan and Walter Benjamin becoming more prominent in Section II. Key technologies in this analysis include mechanization (mechanical reproducibility), the Williamson amplification circuit, the House of Joy speaker, the dub plate (acetate phonograph) and vinyl record, twin-turntables and the microphone, the magnetic tape recorder, and perhaps most importantly, the multi-track recorder and interface (the multi-track mixing-board).

Algorithms exist which will perform independent transformations on frequency or duration of a digital audio signal. These processes have different results different types of audio signals. A comparative study of granular and phase vocoder algorithms, implementation, and their respective effects on audio signals was made to determine which algorithm is best suited to a particular type of audio signal.

No description available.

This thesis is concerned with an examination of the reliability of R-DAT recording media in regard to professional and archival applications. Four brands of R-DAT tapes were subjected to mechanical stress, environmental stress, and a combination of both the mechanical and environmental stresses. Data generated from these tests were analyzed objectively, subjectively evaluated, and subsequently compared. Findings showed that in the majority of cases, the subjective evaluation results corroborate the objective measurements. The study concludes that only one brand of tape exhibited no deterioration of data, while the other three brands failed at various points throughout the testing.

This thesis presents high-level strategies for controlling existing audio morphing algorithms. Prominent methods for spectrally representing audio data in the context of audio morphing are surveyed, and techniques involving manipulating and interpolating additive models are explored. Additionally, a combination of past audio-morphing approaches (sinusoidal morphing techniques as well as techniques taken from an MFCC approach) are employed in order to facilitate any type of sonic input. The system presented herein is meant to accommodate a wide range of inputs, and will adjust the impact of a high-level feature depending on the type and quality of the input.* / *This dissertation is a compound document (contains both a paper copy and a CD as part of the dissertation).