This thesis investigates a fuzzy syntactic approach to the interpretation of noisy two-dimensional images. This approach is based on a modification of the attributed graph grammar formalism to utilise fuzzy membership functions in the applicability predicates. As far as we are aware, this represents the first such modification of graph grammars. Furthermore, we develop a method for programming the resultant fuzzy attributed graph grammars through the use of non-deterministic control diagrams. To do this, we modify the standard programming mechanism to allow it to cope with the fuzzy certainty values associated with productions in our grammar. Our objective was to develop a flexible framework which can be used for the recognition of a wide variety of image classes, and which is adept at dealing with noise in these images. Programmed graph grammars are specifically chosen for the ease with which they allow one to specify a new two-dimensional image class. We implement a prototype system for Optical Music Recognition using our framework. This system allows us to test the capabilities of the framework for coping with noise in the context of handwritten music score recognition. Preliminary results from the prototype system show that the framework copes well with noisy images.

Digital audio networks are becoming the preferred solution for the interconnection of professional audio devices. Prominent amongst their advantages are: reduced noise interference, signal multiplexing, and a reduction in the number of cables connecting networked devices. In the context of professional audio, digital networks have been used to connect devices including: mixers, effects units, preamplifiers, breakout boxes, computers, monitoring controllers, and synthesizers. Such networks are governed by protocols that define the connection management rocedures, and device synchronization processes of devices that conform to the protocols. A wide range of digital audio network control protocols exist, each defining specific hardware requirements of devices that conform to them. Device parameter control is achieved by sending a protocol message that indicates the target parameter, and the action that should be performed on the parameter. Typically, a device will conform to only one protocol. By implication, only devices that conform to a specific protocol can communicate with each other, and only a controller that conforms to the protocol can control such devices. This results in the isolation of devices that conform to disparate protocols, since devices of different protocols cannot communicate with each other. This is currently a challenge in the professional music industry, particularly where digital networks are used for audio device control. This investigation seeks to resolve the issue of interoperability between professional audio devices that conform to different digital audio network protocols. This thesis proposes the use of a proxy that allows for the translation of protocol messages, as a solution to the interoperability problem. The proxy abstracts devices of one protocol in terms of another, hence allowing all the networked devices to appear as conforming to the same protocol. The proxy receives messages on behalf of the abstracted device, and then fulfills them in accordance with the protocol that the abstracted device conforms to. Any number of protocol devices can be abstracted within such a proxy. This has the added advantage of allowing a common controller to control devices that conform to the different protocols.

Performance degradation -- Training from natural vs. automated voice.

No description available.

No description available.

No description available.

Despite the lack of formal guidelines, synthetic speech displays are used in a growing variety of applications. Telephone information systems permitting human-computer interaction from remote locations are an especially popular implementation of computer-generated speech. Currently, human factors research is needed to specify design characteristics providing usable telephone information systems as defined by task performance and user ratings. Previous research used nonintegrated tasks such as transcription of phonetic syllables, words, or sentences to assess task performance or user preference differences. This study used a computer-driven telephone information system as a real-time, human-computer interface to simulate applications where synthetic speech is used to access data. Subjects used a telephone keypad to navigate through an automated, department store database to locate and transcribe specific information messages. Because speech provides a sequential and transient information display, users may have difficulty navigating through auditory databases. One issue investigated in this study was whether use of alternating male and female voices to code different levels in the database hierarchy would improve user search performance. Other issues investigated were basic intelligibility of these male and female voices as influenced by different levels of speech rate. All factors were assessed as functions of search or transcription task performance and user preference. Analysis of transcription accuracy, search efficiency and time, and subjective ratings revealed an overall significant effect of speech rate on all groups of measures but no significant effects for voice type or coding scheme. Results were used to recommend design guidelines for developing speech displays for telephone information systems. / Master of Science

Recent improvements in speech technology have made synthetic speech a viable I/O alternative. However, little research has focused on optimizing the various speech parameters which influence system performance. This study examined the effects of speech rate, message repetition, and the placement of information in a message. Briefly, subjects heard messages generated by a speech synthesizer and were asked to transcribe what they had heard. After entering each transcription, subjects rated the perceived difficulty of the preceding message, and how confident they were of their response. The accuracy of their response, system response time, and response latency were recorded. Transcription accuracy was best for messages spoken at 150 or 180 wpm and for messages repeated either twice or three times. Words at the end of messages were transcribed more accurately than words at the beginning of messages. Response latencies were fastest at 180 wpm with 3 repetitions and rose as the number of repetitions decreased. System response times were shortest when a message was repeated only once. The subjective certainty and difficulty ratings indicated that subjects were aware of errors when incorrectly transcribing a message. These results suggest that a) message rates should lie below 210 wpm, b) a repeat feature should be included in speech interface designs, and c) important information should be contained at the end of messages. / Master of Science

Thesis (MScEng (Electrical and Electronic Engineering))--Stellenbosch University, 2008. / Automatic bird song recognition and transcription is a relatively new field. Reliable automatic recognition systems would be of great benefit to further research in ornithology and conservation, as well as commercially in the very large birdwatching subculture. This study investigated the use of Hidden Markov Models and duration modelling for bird call recognition. Through use of more accurate duration modelling, very promising results were achieved with feature vectors consisting of only pitch and volume. An accuracy of 51% was achieved for 47 calls from 39 birds, with the models typically trained from only one or two specimens. The ALS pitch tracking algorithm was adapted to bird song to extract the pitch. Bird song synthesis was employed to subjectively evaluate the features. Compounded Selfloop Duration Modelling was developed as an alternative duration modelling technique. For long durations, this technique can be more computationally efficient than Ferguson stacks. The application of approximate string matching to bird song was also briefly considered.

This thesis investigates the feasiblity of using a new microcontroller architecture, the XMOS XS1, in the research and development of control standards for audio distribution networks. This investigation is conducted in the context of an emerging audio distribution network standard, Ethernet Audio/Video Bridging (`Ethernet AVB'), and an emerging audio control standard, AES-64. The thesis describes these emerging standards, the XMOS XS1 architecture (including its associated programming language, XC), and the open-source implementation of an Ethernet AVB streaming audio device based on the XMOS XS1 architecture. It is shown how the XMOS XS1 architecture and its associated features, focusing on the XC language's mechanisms for concurrency, event-driven programming, and integration of C software modules, enable a powerful implementation of the AES-64 control standard. Feasibility is demonstrated by the implementation of an AES-64 protocol stack and its integration into an XMOS XS1-based Ethernet AVB streaming audio device, providing control of Ethernet AVB features and audio hardware, as well as implementations of advanced AES-64 control mechanisms. It is demonstrated that the XMOS XS1 architecture is a compelling platform for the development of audio control standards, and has enabled the implementation of AES-64 connection management and control over standards-compliant Ethernet AVB streaming audio devices where no such implementation previously existed. The research additionally describes a linear design method for applications based on the XMOS XS1 architecture, and provides a baseline implementation reference for the AES-64 control standard where none previously existed.