Dissertations / Theses on the topic 'Magnitude spectrum'

Speech is the dominant mode of communication between humans; simple to learn, easy to use and integral for modern life. Given the importance of speech, development of a human-machine speech interface has been greatly anticipated. This challenging task is encapsulated in the digital speech processing research field. In this dissertation, two specific areas of research are considered: 1) the use of short-time Fourier spectral phase in digital speech processing and 2) use of the minimum mean square error spectral energy estimator for environment-robust automatic speech recognition. In speech processing and modelling, the short-time Fourier spectral phase has been considered of minor importance. This is because classic psychoacoustic experiments have shown speech intelligibility to be closely related to short-time Fourier spectral magnitude. Given this result, it is unsurprising that the majority of speech processing literature has involved exploitation of the short-time magnitude spectrum. Despite this, recent studies have shown useful information can be extracted from the spectral phase of speech. As a result, it is now known that spectral phase possesses much of the same intelligibility information as spectral magnitude. It is this avenue of research that is explored in greater detail within this dissertation. In particular, we investigate two phase derived quantities – the short-time instantaneous frequency spectrum and the short-time group delay spectrum. The properties of both spectra are investigated mathematically and empirically, identifying the relationship between known speech features and the underlying phase spectrum. We continue the investigation by examining two related quantities – the instantaneous frequency deviation and the group delay deviation. As a result of this research, two novel phase-based spectral representations are proposed, both of which show a high degree information applicable to speech processing.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
Griffith School of Engineering
Science, Environment, Engineering and Technology
Full Text

Majority of speech processing algorithms that employ the short-time Fourier transform process the short-time magnitude spectrum, while either discarding the short-time phase spectrum or leaving it unchanged. This is in-part due to a long-standing belief among speech researchers that the short-time phase spectrum, computed over small analysis window durations of 20–40 ms, contains little useful information and is thus (mostly) unimportant for speech processing (though it is accepted that the phase spectrum does contribute to some extent to naturalness and quality aspects of speech). The above belief has been supported by numerous studies presented in the literature. Results of recent speech perception experiments suggest, however, that the phase spectrum (at small analysis window durations of 20–40 ms) does contain significant amount of useful information, provided that the analysis window function is carefully selected. It was reported that the use of non-tapered analysis windows functions (such as the rectangular window) significantly improves intelligibility of the phase spectrum. This improvement was attributed to the spectral characteristics of the non-tapered analysis windows and—in particular—to their low spectral dynamic range. The main aim of the research presented in this dissertation is to further examine the importance of the short-time phase spectrum for human speech perception. It is hoped that results of such an examination can provide an incentive for further research in this direction. Three studies that investigate the usefulness of the phase spectrum for human speech perception are presented in this thesis. These studies employ human listening tests to explore the importance of the phase spectrum for speech intelligibility, speaker dependent speech information and speech quality. In each of these studies the effect of the spectral dynamic range of an analysis window function is systematically examined.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
Griffith School of Engineering
Science, Environment, Engineering and Technology
Full Text

Incorporating information from the short-time phase spectrum into a feature set for automatic speech recognition (ASR) may possibly serve to improve recognition accuracy. Currently, however, it is common practice to discard this information in favour of features that are derived purely from the short-time magnitude spectrum. There are two reasons for this: 1) the results of some well-known human listening experiments have indicated that the short-time phase spectrum conveys a negligible amount of intelligibility at the small window durations of 20-40 ms used for ASR spectral analysis, and 2) using the short-time phase spectrum directly for ASR has proven di?cult from a signal processing viewpoint, due to phase-wrapping and other problems. In this thesis, we explore the possibility of using short-time phase spectrum information for ASR by considering the two points mentioned above. To address the ?rst point, we conduct our own set of human listening experiments. Contrary to previous studies, our results indicate that the short-time phase spectrum can indeed contribute signi?cantly to speech intelligibility over small window durations of 20-40 ms. Also, the results of these listening experiments, in addition to some ASR experiments, indicate that at least part of this intelligibility may be supplementary to that provided by the short-time magnitude spectrum. To address the second point (i.e., the signal processing di?culties), it may be necessary to transform the short-time phase spectrum into a more physically meaningful representation from which useful features could possibly be extracted. Speci?cally, we investigate the frequency-derivative (or group delay function, GDF) and the time-derivative (or instantaneous frequency distribution, IFD) as potential candidates for this intermediate representation. We have performed various experiments which show that the GDF and IFD may be useful for ASR. We conduct several ASR experiments to test a feature set derived from the GDF. We ?nd that, in most cases, these features perform worse than the standard MFCC features. Therefore, we suggest that a short-time phase spectrum feature set may ultimately be derived from a concatenation of information from both the GDF and IFD representations. For best performance, the feature set may also need to be concatenated with short-time magnitude spectrum information. Further to addressing the two aforementioned points, we also discuss a number of other speech applications in which the short-time phase spectrum has proven to be very useful. We believe that an appreciation for how the short-time phase spectrum has been used for other tasks, in addition to the results of our research, will provoke fellow researchers to also investigate its potential for use in ASR.

Incorporating information from the short-time phase spectrum into a feature set for automatic speech recognition (ASR) may possibly serve to improve recognition accuracy. Currently, however, it is common practice to discard this information in favour of features that are derived purely from the short-time magnitude spectrum. There are two reasons for this: 1) the results of some well-known human listening experiments have indicated that the short-time phase spectrum conveys a negligible amount of intelligibility at the small window durations of 20-40 ms used for ASR spectral analysis, and 2) using the short-time phase spectrum directly for ASR has proven di?cult from a signal processing viewpoint, due to phase-wrapping and other problems. In this thesis, we explore the possibility of using short-time phase spectrum information for ASR by considering the two points mentioned above. To address the ?rst point, we conduct our own set of human listening experiments. Contrary to previous studies, our results indicate that the short-time phase spectrum can indeed contribute signi?cantly to speech intelligibility over small window durations of 20-40 ms. Also, the results of these listening experiments, in addition to some ASR experiments, indicate that at least part of this intelligibility may be supplementary to that provided by the short-time magnitude spectrum. To address the second point (i.e., the signal processing di?culties), it may be necessary to transform the short-time phase spectrum into a more physically meaningful representation from which useful features could possibly be extracted. Speci?cally, we investigate the frequency-derivative (or group delay function, GDF) and the time-derivative (or instantaneous frequency distribution, IFD) as potential candidates for this intermediate representation. We have performed various experiments which show that the GDF and IFD may be useful for ASR. We conduct several ASR experiments to test a feature set derived from the GDF. We ?nd that, in most cases, these features perform worse than the standard MFCC features. Therefore, we suggest that a short-time phase spectrum feature set may ultimately be derived from a concatenation of information from both the GDF and IFD representations. For best performance, the feature set may also need to be concatenated with short-time magnitude spectrum information. Further to addressing the two aforementioned points, we also discuss a number of other speech applications in which the short-time phase spectrum has proven to be very useful. We believe that an appreciation for how the short-time phase spectrum has been used for other tasks, in addition to the results of our research, will provoke fellow researchers to also investigate its potential for use in ASR.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Microelectronic Engineering
Full Text

Análise de idades e metalicidades de aglomerados estelares nas Nuvens de Magalhães traz informação para estudos sobre a evolução química e dinâmica das Nuvens. Usando-as como calibradores de modelos de populações estelares simples, esse tipo de análise é útil também para o estudo de outras galáxias. Um dos objetivos deste trabalho é derivar idades e metalicidades a partir de espectros integrados de 14 aglomerados na Pequena Nuvem de Magalhães. Busca-se o entendimento, em particular, dos aglomerados de idade intermediária/velha. A metodologia usada para isso é baseada em ajuste de espectro completo dos espectros integrados dos aglomerados, comparando-os a três bases de modelos de populações estelares simples; são usados dois códigos disponíveis na literatura para efetuar tais comparações. Desse modo, são identicados 9 aglomerados como de idade intermediária/velha e os outros 5, jovens. Destacam-se os resultados para os aglomerados com idade intermediária/velha recém identicadas: HW 1, NGC 152, Lindsay 3, 11 e 113. São conrmadas também as idades velhas de NGC 361, 419 e Kron 3 e do bem conhecido e mais velho aglomerado da Pequena Nuvem, NGC 121. Outro objetivo é determinar parâmetros físicos autoconsistentes (idade, metalicidade, distância e avermelhamento) para 7 aglomerados relativamente pouco estudados da Pequena Nuvem, com idades entre ~ 0.5 e 5 Ganos. Para isso são usadas ferramentas estatísticas que comparam CMDs modelados com os observados. Diferentemente de um ajuste visual de isócrona, essa abordagem oferece um critério objetivo e inequívoco para estabelecer quais são os CMDs sintéticos que melhor ajustam o CMD observado. Resultados preliminares mostram a eciência deste método, que determina log(idade), com incerteza de 0.10 e [Fe/H], com 0.20. Isso pode introduzir novos e importantes limites na relação idade-metalicidade da Pequena Nuvem, que é signicativamente mais complexa e menos estudada que a da Grande Nuvem.
Analysis of age and metallicity of star clusters in Magellanic Clouds brings information to studies on the chemical evolution and dynamics of the Clouds. By using them as calibrators of single stellar populations models, this type of analysis is also useful for the study of other galaxies. One of the aims of this work is to derive ages and metallicities from integrated spectra of 14 clusters in the Small Magellanic Cloud. We seek for the understanding, in particular, of the intermediate/old age clusters. The method used for this purpose is based on full spectrum tting of integrated spectra of the clusters, comparing them to three sets of single stellar populations; two codes available in the literature are used to do such comparisons. Thus 9 clusters are identied as intermediate/old age clusters and 5 other as young ones. Noteworthy are the results for the clusters with intermediate/old age newly identied: HW 1, NGC 152, Lindsay 3, 11 and 113. We also conrm the old ages of NGC 361, 419 and Kron 3, and of the well-known oldest cluster of the Small Cloud, NGC 121. Another aim is to determine self-consistent physical parameters (age, metallicity, distance and reddening) for 7 relatively unstudied stellar clusters in the Small Magellanic Cloud, with ages between 0.5 and 5 Gyr. We use statistical tools to compare synthetic and observed CMDs. Dierently of a visual isochrone t, this approach oers ob jective and unambiguous criteria to establish which are the synthetic CMDs that best ts the observed CMD. Preliminary results show the eciency of this method, which determines log(age), with an uncertainty of 0.10 and [Fe/H], with 0.20. This can introduce new and important constraints in the age-metallicity relation for the Small Cloud, which is signicantly more complex and less studied than the one of the Large Magellanic Cloud.

Thesis (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1999.
Includes bibliographical references (p. 43-44).
by Charles Kasimer Sestok, IV.
S.M.

The primary goal of this paper is to demonstrate the application of a relatively esoteric and interdisciplinary technique, called spectral analysis, to dyadic social sciences data. Spectral analysis is an analytical and statistical technique, commonly used in engineering, that allows times series data to be analyzed for the presence of significant regular/periodic fluctuations/oscillations. These periodic fluctuations are reflected in the frequency domain as amplitude or energy peaks at certain frequencies. Furthermore, a Magnitude Squared Coherence analysis may be used to interrogate more than one time series concurrently in order to establish the degree of frequency domain correlation between the two series, as well to establish the phase (lead/lag) relationship between the coherent frequency components. In order to demonstrate the application of spectral analysis, the current study utilizes a secondary dyadic dataset comprising 30 daily reports of perceived sexual desire for 65 couples. The secondary goal of this paper is to establish a) whether there is significant periodic fluctuation in perceived levels of sexual desire for men and/or women, and at which specific frequencies, and b) how much correlation or `cross-spectral coherence' there is between partners' sexual desire within the dyads, and c) what the phase lead-lag relationship is between the partners at any of the identified frequency components. Sexual desire was found to have significant periodic components for both men and women, with a fluctuation of once per month being the most common frequency component across the groups of individuals under analysis. Mathematical models are presented in order to describe and illustrate these principal fluctuations. Partners in couples, on average, were found to fluctuate together at a number of identified frequencies, and the phase lead/lag relationships of these frequencies are presented.

By studying the frequency domain of seismic signals generated by earthquakes, the source parameters can be recovered, i.e., the seismic moment (M0) and the stress drop (Δσ). This method is an advantage especially since if the source parameters are calculated from the time domain a full waveform inversion is needed, therefore this procedure facilitates the computation. Besides, the moment magnitude (Mw) can be calculated from the seismic moment and, in turn, the local magnitude (ML) can be obtained by using an algorithm that matches different ranges of moment magnitude with their corresponding local magnitude. In the present thesis, small to moderate earthquakes in Northern Sweden have been used to develop a code that calculates the source parameters through the fitting of five different spectral models and, this way, discerns which model obtains the best determination of the parameters. These models have been chosen in a way that we can also extract information about the attenuation. The different models are; the Brune spectral model, Boatwright spectral model, Boatwright spectral model with a fixed fall-off rate, a general form of the spectral model with quality factor equal to 1000 and a general form of the spectral model with quality factor equal to 600. Among these models, the Boatwright model with fixed fall-off rate equal to 2, has been found to give the best fit to the data used in this thesis. This might be due to the regional conditions which are the low attenuation in the crust of northern Fennoscandia and the short hypocentral distances of the studied earthquakes. The earthquakes studied in the present thesis have shown a range of magnitudes from ML 4.2 to -0.2 with radius of an assumed circular fault ranging from 269 m to 66 m.

Speech signal processing for voice communications can be characterised in terms of silence compression, noise reduction, and speech compression. The limit in the channel bandwidth of voice communication systems requires efficient compression of speech and silence signals while retaining the voice quality. Silence compression by means of both voice activity detection (VAD) and comfort noise generation could present transparent speech-quality while substantially lowering the transmission bit-rate, since pause regions between talk spurts do not include any voice information. Thus, this thesis proposes smoothed likelihood ratio-based VAD, designed on the basis of a behavioural analysis and improvement of a statistical model-based voice activity detector. Input speech could exhibit noisy signals, which could make the voice communication fatiguing and less intelligible. This task can be alleviated by noise reduction as a preprocessor for speech coding. Noise characteristics in speech enhancement are adapted typically during the pause regions classified by a voice activity detector. However, VAD errors could lead to over- or under- estimation of the noise statistics. Thus, this thesis proposes mixed decision-based noise adaptation based on a integration of soft and hard decision-based methods, defined with the speech presence uncertainty and VAD result, respectively. At low bit-rate speech coding, the sinusoidal model has been widely applied because of its good nature exploiting the phase redundancy of speech signals. Its performance, however, can be severely smeared by mis-estimation of the pitch. Thus, this thesis proposes a robust pitch estimation technique based on the autocorrelation of spectral amplitudes. Another important parameter in sinusoidal speech coding is the spectral magnitude of the LP-residual signal. It is, however, not easy to directly quantise the magnitudes because the dimensions of the spectral vectors are variable from frame to frame depending on the pitch. To alleviate this problem, this thesis proposes mel-scale-based dimension conversion, which converts the spectral vectors to a fixed dimension based on mel-scale warping. A predictive coding scheme is also employed in order to exploit the inter-frame redundancy between the spectral vectors. Experimental results show that each proposed technique is suitable for enhancing speech quality for voice communications. Furthermore, an improved speech coder incorporating the proposed techniques is developed. The vocoder gives speech quality comparable to TIA/EIA IS-127 for noisy speech whilst operating at lower than half the bit-rate of the reference coder. Key words: voice activity detection, speech enhancement, pitch, spectral magnitude quantisation, low bit-rate coding.

The problem of sidelobes surrounding a peak in a Fourier transform spectrum is alleviated by apodization. This is performed by multiplying the time-domain function by a window function. A systematic study of the effects of window functions on damped time-domain signals is made by examining the resulting lineshapes at specific dynamic ranges for both the absorption and magnitude modes. A symmetrical window is shown to be effective for the magnitude mode, and half of the symmetrical shape is better for the absorption mode. Selection of a recommended window is based on the required dynamic range. For an increasing dynamic range, the Noest-Kort and Norton-Beer F3, Filler E0.20, and Kaiser-Bessel are efficient for the absorption mode; and the Hamming, 3-term Blackman-Harris and Kaiser-Bessel work for the magnitude mode. Sidelobes are often eliminated at the expense of SNR and/or resolution, therefore these factors are also examined. All of the recommended windows show sufficient SNRs except for the Noest-Kort. The apodized absorption spectra are well resolved, with a 10% valley as the criterion for resolution. The magnitude spectra do not display a simple pattern, and also show a phase dependence; however these are explained by the absorption and dispersion components. These findings lend themselves to various applications.
Science, Faculty of
Chemistry, Department of
Graduate

The doctoral thesis deals with spectrum sensing and subsequent use of the frequency spectrum by multicarrier communication system, which parameters are set on the basis of the optimization technique. Adaptation settings can be made with respect to several requirements as well as state and occupancy of individual communication channels. The system, which is characterized above is often referred as cognitive radio. Equipments operating on cognitive radio principles will be widely used in the near future, because of frequency spectrum limitation. One of the main contributions of the work is the novel usage of the Kolmogorov – Smirnov statistical test as an alternative detection of primary user signal presence. The new fitness function for Particle Swarm Optimization (PSO) has been introduced and the Error Vector Magnitude (EVM) parameter has been used in the adaptive greedy algorithm and PSO optimization. The dissertation thesis also incorporates information about the reliability of the frequency spectrum sensing in the modified greedy algorithm. The proposed methods are verified by the simulations and the frequency domain energy detection is implemented on the development board with FPGA.

Les interféromètres stellaires à très haute résolution angulaire sont des dispositifs performants qui permettent une étude approfondie de notre Univers. Cette thèse décrit un nouvel instrument, baptisé interféromètre à somme de fréquences, dédié à l’analyse de la cohérence spatio-temporelle de rayonnements infrarouges autour de 1550nm en les transposant vers les longueurs d’onde visibles autour de 630nm. Cette transposition fréquentielle est réalisée au moyen de processus de somme de fréquences, alimentés par une source laser à 1064nm, et générés dans les cristaux de PPLN mis en place sur chacun des bras de l’instrument. La première partie de ce manuscrit donne les éléments théoriques nécessaires à la compréhension des travaux présentés. La deuxième partie concerne la conception, la réalisation et la caractérisation d’un interféromètre à somme de fréquences adapté à l’analyse de sources infrarouges spectralement étendues. La troisième partie présente la caractérisation expérimentale de la cohérence temporelle d’une source infrarouge large bande spectrale. Elle se solde par la démonstration expérimentale de l’analyse de la cohérence spatiale d’une source thermique artificielle en régime de comptage de photons. La dernière partie traite de la possibilité d’étendre la bande spectrale de fonctionnement de l’interféromètre et démontre l’apparition un effet de compression spectrale en fréquences dans les cristaux de PPLN. Elle se conclut par le projet de mise en place de l’instrument sur le site du réseau de télescope CHARA et fournit une estimation de la magnitude limite accessible en combinant ce dispositif à différentes configurations de l’interféromètre
Stellar interferometers dedicated to high angular resolution imaging are powerful instruments that allow an extensive study of our Universe. This thesis describes a new instrument, called sum frequency generation interferometer, dedicated to the spatio-temporal coherence analysis of infrared radiations in the visible domain around 630nm thanks to frequency shift processes. This frequency shift is achieved through a sum frequency generation process, supplied by a laser source at 1064 nm, in the PPLN crystals placed on each arm of the interferometer. The first part of this thesis provides the theoretical elements needed for the understanding of the present work. The second part concerns the design, the realization and the characterization of a sum frequency generation interferometer suitable for the analysis of broadband infrared sources. The third part presents the experimental temporal coherence analysis of such a source. It ends by the experimental demonstration of the spatial coherence analysis of an artificial thermal source in the photon counting regime. The last part deals with the possibility to extend the spectral bandwidth of the interferometer and shows the appearance of a frequency spectral compression effect through the PPLN crystals. It concludes with the proposed implementation of a sum frequency generation interferometer on the CHARA telescope array and provides an estimation of the limiting magnitude reached by combining this instrument to different configurations of the interferometer

Denna rapport täcker arbetsprocessen bakom att skapa en spektrumanalysator. Läsaren kommer att få läsa om den valda metoden men även alternativa metoder. Utöver detta kommer även de teoretiska delarna bakom varje moment att undersökas samt jämföras med potentiella alternativa lösningar. Projektet har utförts på uppdrag av KA Automation. Syftet med projektet var att skapa en basplattform för analys av ljudfrekvenser. Målet med detta var att kunna identifiera ljudegenskaper i form av frekvenserna hos exempelvis servomotorer i vattenpumpar. Tanken var att i ett senare utvecklingsskede kunna identifiera om och när nya frekvenser dykt upp i ljudprofilen vilket i sådana fall kan resultera i att motorn är i behov av service. Basplattformen är uppbyggd med hjälp av C# och ljudbehandlingsbiblioteket NAudio. Från resultatet kan slutsatsen dras att detta program kan analysera ljud och visa de olika frekvensernas styrka och därmed är en lämplig basplattform för vidareutveckling.
This report will cover the work process behind creating a spectrum analyzer. The reader will be able to read about the chosen method but also the alternative methods. Apart from this the theoretical parts behind every moment will also be covered and compared to potential alternative solutions. The project has been carried out on behalf of KA Automation. The purpose of the project was to create a base for analyzing sound frequencies. The goal was to be able to identify sound properties in the form of frequencies in servo motors in for example water pumps. The idea was to be able to in a later development stage be able to identify when new frequencies have entered the audio profile which might result in the motor to be in need of service. The base is created with the help of C# and the sound library NAudio. From the result one can conclude that this program can analyze sound and display the magnitude of its frequency components and is therefore a suitable base for future development.

The aim of this thesis was to develop user interface in LabVIEW to make typical measurements with spectral analyzer HP89410A There is introduced the theory of operation of an analogue heterodyne and a digital FFT spectral analyzer. It is explained the background of the Fast Fourier Transform. There are discussed the key settings in measuring with FTT analyzer, like window function, bandwidth, number of frequency points, or the averaging. Further is described the program, which was developed for measuring frequency characteristic by white noise and by stepped measurement. It was also made a module for measuring Signal to Noise Ratio and module for Total Harmonic Distortion. In the last part of this thesis are shown the results of processed exemplar measurements.

Pour reconstituer l'histoire des grandes structures de l'Univers, il faut pouvoir observer les caractéristiques de l'Univers à différentes époques du passé et comprendre les mécanismes physiques impliqués dans l'évolution des galaxies. Grâce aux progrés de l'instrumentation réalisé ces dix dernières années, la communauté a réalisé de larges programmes d'observations, de plus en plus profonds et à différentes longueurs d'onde. L'un d'eux qui est à la base de ce travail, est le plus large est profond à la fois à l'heure actuelle et a été réalisé au CFHT. Grâce au code Hyperz, nous avons calculé les redshifts photométriques des galaxies du CFHTLS "Deep" et nous avons éudié l'évolution de la relation couleur-magnitude de nos échantillons jusqu'à z~1. 2. Nous avons également étudié le role de l'environnement sur les propriétés des galaxies grace à un indicateur de densité locale projetée. Nous avons montré que la distribution en couleur des galaxies est bimodale dans tout les intervalles de magnitudes et de redshifts étudiés. Nous avons observé une forte évolution de cette distribution encouleur, en fonction du redshift et de la luminosité, avec une évolutionplus douce en fonction de la densité locale. Les galaxies étaient significativement plus bleues dans le passée et au moins 50% de la population rouge brillante à z~1. 2 était déjà en place, ce qui indique une époque d'assemblage précoce. L'environnement semble jouer un role à bas redshift jusqu'à z~0. 4, commecela a été observé dans le SDSS. La population rouge et brillante est surreprésentée dans les régions très denses, alors que l'on trouve une importante population bleue peu brillante dans les régions vides. Cette tendance s'inverse à plus grand redshift, où c'est la luminosité qui devient le paramétre qui influence le plus les couleurs moyennes des galaxies. Les relations couleur-magnitude-environnement des populations observées dans l'Univers local étaient déjà en place à un redshift d'environ z~0. 6. Nous avons réalisé des tests qui nous ont permis de valider notre méhode de calcul des masses stellaires des galaxies. .
In order to understand the evolution of the big structures in the Univers and the physical mecanisms implicated, we need to observe the caracteristics of galaxies at different redshift in the past. Thanks to progress in instrumentation technics made in the last ten years, the community has realised wide and deep observation programs, at different wavelenghts. One of them, realised at the CFHT, is both the widest and deepest achieved until now and has been studied in this thesis: the CFHTLS Deep. We have selected complete samples of galaxies at different redshift bin up to z~1. 2 in the CFHTLS Deep catalogs. We hace computed photometric redshifts for every galaxies with the code Hyperz, and their absolute magnitude in the five filters (u*g'r'i'z'). Thus, we have been able to study the evolution of the color-magnitude relations of our samples. We have also studied the role of environment on the galaxies properties thanks to a projected local density estimator calculated for every galaxies in the CFHTLS Deep four fields. We have shown that the color distribution of galaxies is bimodal in every magnitude and redshift bins studied. We have observed a strong evolution of the color distribution with redshift and luminosity, together with a midl evolution with local density. The color-magnitude-environment relations of the galaxies populations observed in local Univers were already built at z~0. 6. Galaxies were significantly bluer in the past and almost 50% of the bright red population were already built at z~1. 2, which indicates an early epoch of assembly. Beyond z~0. 8, an important blue and bright population appears. .

Thesis (Ph.D.)--Electrical and Computer Engineering, Georgia Institute of Technology, 2009.
Committee Chair: James Stevenson Kenney; Committee Member: Gregory David Durgin; Committee Member: Madhavan Swaminathan

La sélection de caractéristiques acoustiques appropriées est essentielle dans tout système de traitement de la parole. Pendant près de 40 ans, la parole a été généralement considérée comme une séquence de signaux quasi-stables (voyelles) séparés par des transitions (consonnes). Bien qu‟un grand nombre d'études documentent clairement l'importance de la coarticulation, et révèlent que les cibles articulatoires et acoustiques ne sont pas indépendantes du contexte, l‟hypothèse que chaque voyelle présente une cible acoustique qui peut être spécifiée d'une manière indépendante du contexte reste très répandue. Ce point de vue implique des limitations fortes. Il est bien connu que les fréquences de formants sont des caractéristiques acoustiques qui présentent un lien évident avec la production de la parole, et qui peuvent participer à la distinction perceptive entre les voyelles. Par conséquent, les voyelles sont généralement décrites avec des configurations articulatoires statiques représentées par des cibles dans l'espace acoustique, généralement par les fréquences des formants correspondants, représentées dans les plans F1-F2 et F2-F3. Les consonnes occlusives peuvent être décrites en termes de point d'articulation, représentés par locus (ou locus équations) dans le plan acoustique. Mais les trajectoires des fréquences de formants dans la parole fluide présentent rarement un état d'équilibre pour chaque voyelle. Elles varient avec le locuteur, l'environnement consonantique (co-articulation) et le débit de parole (relative à un continuum entre hypo et hyper-articulation). En vue des limites inhérentes aux approches statiques, la démarche adoptée ici consiste à étudier les transitions entre les voyelles et les consonnes (V1V2 et V1CV2) d‟un point de vue dynamique
Speech plays a vital role in human communication. Selection of relevant acoustic speech features is key to in the design of any system using speech processing. For some 40 years, speech was typically considered as a sequence of quasi-stable portions of signal (vowels) separated by transitions (consonants). Despite a wealth of studies that clearly document the importance of coarticulation, and reveal that articulatory and acoustic targets are not context-independent, the view that each vowel has an acoustic target that can be specified in a context-independent manner remains widespread. This point of view entails strong limitations. It is well known that formant frequencies are acoustic characteristics that bear a clear relationship with speech production, and that can distinguish among vowels. Therefore, vowels are generally described with static articulatory configurations represented by targets in the acoustic space, typically by formant frequencies in F1-F2 and F2-F3 planes. Plosive consonants can be described in terms of places of articulation, represented by locus or locus equations in an acoustic plane. But formant frequencies trajectories in fluent speech rarely display a steady state for each vowel. They vary with speaker, consonantal environment (co-articulation) and speaking rate (relating to continuum between hypo- and hyper-articulation). In view of inherent limitations of static approaches, the approach adopted here consists in studying both vowels and consonants from a dynamic point of view.Firstly we studied the effects of the impulse response at the beginning, at the end and during transitions of the signal both in the speech signal and at the perception level. Variations of the phases of the components were then examined. Results show that the effects of these parameters can be observed in spectrograms. Crucially, the amplitudes of the spectral components distinguished under the approach advocated here are sufficient for perceptual discrimination. From this result, for all speech analysis, we only focus on amplitude domain, deliberately leaving aside phase information. Next we extent the work to vowel-consonant-vowel perception from a dynamic point of view. These perceptual results, together with those obtained earlier by Carré (2009a), show that vowel-to-vowel and vowel-consonant-vowel stimuli can be characterized and separated by the direction and rate of the transitions on formant plane, even when absolute frequency values are outside the vowel triangle (i.e. the vowel acoustic space in absolute values).Due to limitations of formant measurements, the dynamic approach needs to develop new tools, based on parameters that can replace formant frequency estimation. Spectral Subband Centroid Frequency (SSCF) features was studied. Comparison with vowel formant frequencies show that SSCFs can replace formant frequencies and act as “pseudo-formant” even during consonant production.On this basis, SSCF is used as a tool to compute dynamic characteristics. We propose a new way to model the dynamic speech features: we called it SSCF Angles. Our analysis work on SSCF Angles were performed on transitions of vowel-to-vowel (V1V2) sequences of both Vietnamese and French. SSCF Angles appear as reliable and robust parameters. For each language, the analysis results show that: (i) SSCF Angles can distinguish V1V2 transitions; (ii) V1V2 and V2V1 have symmetrical properties on the acoustic domain based on SSCF Angles; (iii) SSCF Angles for male and female are fairly similar in the same studied transition of context V1V2; and (iv) they are also more or less invariant for speech rate (normal speech rate and fast one). And finally, these dynamic acoustic speech features are used in Vietnamese automatic speech recognition system with several obtained interesting results

碩士
國立臺北科技大學
電機工程系研究所
102
Speech signals are tend to decrease the speech quality when corrupted by background noises. The main purpose of speech enhancement systems is to reduce the background noise from a noisy speech signal by using both noise estimation algorithm and speech enhancement algorithm, such that produce an enhanced speech signal that has a relatively low speech distortion. In this thesis, we develop two speech enhancement algorithms based on perceptually motivated estimators of the magnitude spectrum together with three noise estimation algorithms. Speech enhancement algorithms include Weighted Euclidean (WE) and Weighted Hyperbolic Cosine (WCOSH). Noise estimation algorithms include Minimum Statistics (MS), Minima Controlled Recursive Averaging (MCRA), and Improved Minima Controlled Recursive Averaging (IMCRA). The experimental results show that compared with the Wiener filter and Minimum Mean-Square Error (MMSE), the perceptually motivated estimators provide better SSNRI, SDI and NRF scores. Among all, the Weighed Cosh method incorporated with MCRA achieves the most significant enhancement performance.

碩士
國立臺灣大學
地質科學研究所
99
We apply the coda-derived source spectrum method to estimate earthquake magnitudes in Taiwan. 76 events occurred during the time period from 1996 to 2010 around west-south Taiwan recorded by BATS (Broadband Array in Taiwan for Seismology) with magnitudes between Mw 3.2 and 6.21 are used in this study. From two horizontal components of waveform data, we extract the coda envelopes at 12 frequency bands between 0.05 and 8.0 Hz. We construct the synthetic coda-envelope by using various empirical distance-frequency-dependent corrections. The derived synthetic coda envelops are compared to the observed coda-envelope to evaluate the dimensionless coda amplitudes. With further considerations on path effects, site effects and source-to-coda transfer functions, the dimensionless coda amplitudes at each frequency band are used to construct the source spectra, which is then used to determine the earthquake magnitudes. The 76 derived coda-magnitudes (MC) are consistent with corresponding moment magnitudes (Mw). Our result also shows the overall deviations between MC and local magnitude (ML) reported from Central Weather Bureau are less than those between Mw and ML. While the implications need further investigation, the results indicate that the broad-band coda information does reconcile, to some extent, the long-lasting discrepancy between the low frequency magnitude (Mw) and the high frequency magnitude (ML) in Taiwan. Moreover, we found out that there is a very stable linear relationship between coda-derived source amplitudes at 2.5 Hz and the ML scale, implying that the source information is well maintained by the high-frequency coda, and this property shows it is potentially very feasible to extend this method to smaller earthquakes in Taiwan.

碩士
國立臺灣大學
地質科學研究所
103
Seismic stations including the Broadband Array in Taiwan for Seismology (BATS) were incorporated within the Central Weather Bureau Seismic Network (CWBSN) since 2012. However, site effects of seismic stations have not been applied for M_L determination that it may affect the magnitude estimation. Moreover, the appearance of relatively low seismicity is observed in 2012 from monthly event counts for〖 M〗_L >2.0. Thus, we would like to examine whether there is a relationship between the updated seismic network and low seismicity or not. In this study, we first collect the earthquake catalog data during the period from 1994/01/01 to 2012/12/31 to calculate the station correction in order to reduce site effect in estimating magnitude. There is a strong correlation between station correction and near-surface geology beneath the station. Stations located on soil sites have high amplifications with negative station corrections. In contrast, stations located on hard rock sites have low amplifications with positive station corrections. After applied the station corrections to revise〖 M〗_L estimation, the monthly event-occurrence rates return inside the range of one standard deviation in 2012 and revealed no low or high anomaly. Therefore, the reason low seismicity in 2012 is that new seismic stations which are installed on rock sites lead to underestimate of the magnitude determination. Next, we investigate the spectral decay parameter kappa (κ) to get more information for site effects of seismic stations. We used 133 earthquakes recorded by 30 borehole seismic array deployed by Central Weather Bureau (CWB) to estimate the κ value. Each vertical array includes two force balance accelerometers at the surface and the borehole. Based on the regression analysis between κ value and hypocentral distance for each surface-borehole station pairs, most of resulting κ0 derived from surface stations are higher than results of borehole stations. These valuable dataset of κ0 values offer an excellent opportunity for us to evaluate the quality factor (Qs) at shallow depths. In comparison with a study of site corrections, these higher κ0 values associated with negative station corrections can correspond to effect of soil conditions. In order to determine the linear relationship between station corrections and κ0 values, we find the co-site stations of TSMIP and CWBSN to estimate κ0 and the regression line. In this study, the station corrections of magnitude and κ0 values both can investigate the site effects beneath the stations. Furthermore, the regression line of these two can be the information for the estimation of earthquake magnitude including borehole stations in the future.

The aim of this thesis is to create collections of examples for the subject Astronomy taught for students of pedagogical faculties, studying this discipline as a part of physics courses. Due to very different mathematical knowledge of students I have chosen typical and not much difficult examples oriented to several branches of astronomy. Each part of examples begins with a self-contained theoretical introduction. The difficulty rises gradually from trivial to more complicated examples. The examples are mainly focused on motions in radial gravitational fields.