La classification de signaux en grande dimension rend nécessaire la sélection d'un petit nombre de structures caractéristiques pour représenter chaque signal. Les approximations non-linéaires donnent lieu à des représentations concises, parce qu'elles s'adaptent à la structure de chaque signal analysé. Leur emploi est prometteur. Une première partie du travail du thèse définit des représentations adaptatives rapides de signaux comme combinaisons linéaires d'atomes extraits d'un dictionnaire de vecteurs. A partir de l'algorithme de Matching Pursuit, plusieurs méthodes itératives sont proposées pour mettre en lumière les structures caractéristiques des signaux sonores. Le Matching Pursuit Harmonique décompose un signal en composantes harmoniques élémentaires. Le Matching Pursuit "Chirpé" extrait les variations de fréquence instantanée en tirant parti d'une analyse fine des crêtes du dictionnaire de Gabor multi-échelle. Les approximations fournies par le Matching Pursuit Haute-résolution préservent les transitoires des signaux analysés, en imposant des contraintes de résolution temporelle. Nous accélérons ces techniques en employant des sous-dictionnaires de maxima locaux. Notre travail est consacré dans un second temps à l'étude de l'"Analyse Discriminante Non-linéaire". Pour classifier des signaux, les méthodes d'Analyse Discriminante Linéaire réduisent la dimension en les projetant sur un sous-espace pré-déterminé. Une projection adaptative, en fonction du signal analysé, extrait de celui-ci des caractéristiques qui lui sont propres. Celles-ci le distinguent et permettent de le classifier efficacement. Nous déterminons la stratégie optimale de projection adaptative pour la classification de bruits gaussiens colorés. Afin de classifier des transitoires, nous explorons enfin une méthode utilisant les maxima du module de la transformée en ondelettes et des arbres de décision. Cette approche permet de surmonter les difficultés liées à l'invariance par translation des signaux à classifier.

Cette thèse tisse des liens entre les domaines de l'indexation audio et de l'analyse de séquences d'images, à travers le problème de l'analyse de l'accompagnement rythmique des signaux de musique. Nous étudions d'abord le problème de la transcription de la piste de batterie d'enregistrements musicaux à partir de la modalité audio seule. Après avoir présenté des pré-traitements permettant d'accentuer cette piste, nous utilisons des techniques d'apprentissage statistique sur une large gamme d'attributs pour réaliser la transcription. Nous introduisons également des approches supervisées et non supervisées pour améliorer la reconnaissance en exploitant la régularité des motifs rythmiques. Nous incorporons ensuite à ce système de transcription l'information visuelle provenant de caméras filmant le batteur. Différentes approches sont présentées pour détecter la position des éléments de la batterie dans la scène et pour associer les régions extraites à des catégories sonores. Des descripteurs d'intensité de mouvement sont ensuite utilisés pour la détection des frappes. Les résultats démontrent la capacité d'une approche multimodale à résoudre certaines des ambiguïtés propres à la transcription audio. Nous étendons enfin nos travaux aux clips vidéos, en tentant de mesurer de quelle manière la musique peut être illustrée par des images. Après avoir présenté ou introduit de nouvelles méthodes de structuration automatique des flux audio et vidéo à différents niveaux, nous définissons des mesures de synchronie sur les structures obtenues. Ces mesures, qui s'avèrent dépendantes du type de document musical, permettent des applications de recherche de musique par l'image.

<p>The effects of using limited precision floating point for intermediate storage in an embedded MP3 decoder are investigated in this thesis. The advantages of using limited precision is that the values need shorter word lengths and thus a smaller memory for storage. </p><p>The official reference decoder was modified so that the effects of different word lengths and algorithms could be examined. Finally, a software and hardware prototype was implemented that uses 16-bit wide memory for intermediate storage. The prototype is classified as a limited accuracy MP3 decoder. Only layer III is supported. The decoder could easily be extended to a full precision MP3 decoder if a corresponding increase in memory usage was accepted.</p>

<p>The thesis project consists on developing an interface for a Nios II processor integrated in a board of Altera (UP3- 2C35F672C6 Cyclone II).</p><p>The main goal is show how the Nios II processor can interact with the other components of the board.The Quartus II software has been used to create to vhdl code of the interfaces, compile it and download it into the board. The Nios II IDE tool is used to build the C/C++ files and download them into the processor.</p><p>It has been prepared an application for the audio codec integrated in the board (Wolfson WM8731 24-bit sigma-delta audio CODEC). The line input of the audio codec receives an analog signal from a laptop, this signal is managed by the control interface of the audio codec. The converters ADCs and DACs are stereo 24-bit sigma delta and they are used with oversampling digital interpolation and decimation filters.</p><p>The digital interface of the audio codec sends the digital signal to the Nios II processor and receives the data from the processor. After building the interfaces for the audio codec and the processor, it has been prepared an application in C++ language for the processor that modifies the volume of the signal.</p><p>The signal come back to the audio codec and it is possible to check the results with headphones or speakers at the line output of the audio codec.</p>

<p>En skillnad i internets utveckling kan ses mellan de visuella och de auditiva elementen, där det auditiva ligger långt efter vad gäller övergripande standarder och användningens utbredning. Vad beror det på? Har skillnader mellan hur vi ser och hör information en betydelse? Denna uppsats undersöker inställningen till ljud över internet med hjälp av en enkät som vanliga internetsurfare med varierande vana av informationssökning över internet har svarat på. Resultaten visar en, inte helt oväntad och inte heller osminkad, skepsis mot ljud över internet.</p>

With the wide-spread adoption of hypermedia systems and the World Wide Web (WWW) in particular, these systems have evolved from simple systems with only textual content to those that incorporate a large content base, which consists of a wide variety of document types. Also, with the increase in the number of users, there has grown a need for these systems to be accessible to a wider range of users. Consequently, the growth of the systems along with the number and variety of users require new presentation and navigation mechanisms for a wider audience. One of the new presentation methods is the audio-only presentation of hypertext content and this research proposes a novel solution to this problem for complex and dynamic systems. The hypothesis is that the proposed Audio Browser is an efficient tool for presenting hypertext in audio format, which will prove to be useful for several applications including browsers for visually-impaired and remote users. The Audio Browser provides audio-only browsing of contents in a Petri-based hypertext system called Context-Aware Trellis (caT). It uses a combination of synthesized speech and pre-recorded speech to allow its user to listen to contents of documents, follow links, and get information about the navigation process. It also has mechanisms for navigating within documents in order to allow users to view contents more quickly.

In the present day, the music becomes more popular due to the following three reasons: (1) the evolution of the MP3 compression technology, (2) the growth of the public platform, and (3) the development of the MP3 portable discs. Most people follow the music to hum or follow the rhythm to tap sometimes. The meanings of a music style may be various if it is explained or felt by different people. Therefore we cannot obtain a very explicit answer if the notation of the music cannot be exactly made. We need some techniques and methods to analyze the music, and obtain some of its embedded information. Tempo and beats are very important elements in the perceptual music. Therefore, tempo estimation and beat tracking are fundamental techniques in automatic audio processing, which are crucial to multimedia applications. In this thesis, we first develop an artificial neural network to classify the music excerpts into the evaluation preference. And then, with the preference classification, we can obtain accurate estimation for tempo and beats, by either Ellis's method or Dixon's method. We test our method with a mixed data set which contains ten music genres extracted from the "ballroom dancer" database. Our experimental results show that the accuracy of our method is higher than that of only one individual Ellis's method or Dixon's method.

<p>Communication tools such as text-based instant messaging, voice and video relay services, real-time video chat and mobile SMS and MMS have successfully been used among Deaf people.&nbsp / Several years of field research with a local Deaf community revealed that disadvantaged South African Deaf people preferred to communicate with both Deaf and hearing peers in South African&nbsp / Sign Language as opposed to text. Synchronous video chat and video relay services provided such opportunities. Both types of services are commonly available in developed regions, but not in&nbsp / developing countries like South Africa. This thesis reports on a workaround approach to design and develop an asynchronous video communication tool that adapted synchronous video codecs&nbsp / to store-and-forward video delivery. This novel asynchronous video tool provided high quality South African Sign Language video chat at the expense of some additional latency. Synchronous video&nbsp / codec adaptation consisted of comparing codecs, and choosing one to optimise in order to minimise latency and preserve video quality. Traditional quality of service metrics only addressed real-time video quality and related services. There was no such standard for asynchronous video communication. Therefore, we also enhanced traditional objective video quality&nbsp / metrics with subjective assessment metrics conducted with the local Deaf community.&nbsp / </p>

A novel framework based on cluster co-occurrence and graph theory for structure discovery is applied to audio to find new types of audio objects which enable the compression of an input signal. These new objects differ from those found in current object coding schemes as their shape is not restricted by any a priori psychoacoustic knowledge. The framework is novel from an application perspective, as it marks the first time that graph theory is applied to audio, and with regards to theoretical developments, as it involves new extensions to the areas of unsupervised learning algorithms and frequent subgraph mining methods. Tests are performed using a corpus of audio files spanning a wide range of sounds. Results show that the framework discovers new types of audio objects which yield average respective overall and relative compression gains of 15.90% and 23.53% while maintaining a very good average audio quality with imperceptible changes.

Stair-climbing is an important skill for promoting independence and activities of daily life and is a key component of rehabilitation therapies for physically disabled children. This thesis describes the design and evaluation of an interactive musical stairs system for children engaged in stair-climbing physical therapies. The achievement of a targeted therapeutic goal, namely, use of reciprocal steps, was significantly increased by 6% [SD=7%] (p=0.007) with the presence of audio feedback. Levels of participant enjoyment and motivation increased as well. This led to the development of an automated system, using inertial sensors to detect initial contact (IC) events each time a child makes a step, to trigger audio feedback. A semi-generic algorithm was designed that was able to detect 96% [SD=3%] of IC events during stair-climbing therapy sessions. This thesis lays the groundwork for future longitudinal research investigating the efficacy of audio feedback in stair-climbing and other rehabilitation therapies as well.