Made available in DSpace on 2014-06-11T19:30:32Z (GMT). No. of bitstreams: 0 Previous issue date: 2009-11-25Bitstream added on 2014-06-13T21:01:20Z : No. of bitstreams: 1 oliveira_vjs_dr_ilha.pdf: 2584742 bytes, checksum: ae7b3113a196a5051a808dbb371dece4 (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / Nesta tese, propõe-se um sintetizador de freqüência baseado em phase locked loops (PLL) usando uma arquitetura que utiliza um dual-path loop filter, constituído de componentes passivos e um integrador digital. A proposta é empregar técnicas digitais, para reduzir o custo da implementação do sintetizador de freqüência, e flexibilizar o projeto do loop filter, para possibilitar que a arquitetura opere em uma faixa de freqüência larga de operação e com redução de tons espúrios. O loop filter digital é constituído de um contador crescente/ decrescente cujo clock é proveniente da amostragem da diferença de fase de entrada. As técnicas digitais aplicadas ao loop filter se baseiam em alterações da operação do contador, em tempos pré-estabelecidos, os quais são controlados digitalmente. Essas técnicas possibilitam reduzir o tempo de estabelecimento do PLL ao mesmo tempo em que problemas de estabilidade são resolvidos. No desenvolvimento da técnica de dual-path foi realizado o estudo de sua estabilidade, primeiramente, considerando a aproximação do PLL para um sistema linear e depois usando controle digital. Nesse estudo foram deduzidas as equações do sistema, no domínio contínuo e discreto, tanto para o projeto da estabilidade, quanto para descrever o comportamento do PLL. A metodologia top-down é usada no projeto do circuito integrado. As simulações em nível de sistema são usadas, primeiramente, para as criações das técnicas e posteriormente para a verificação do seu comportamento, usando modelos calibrados com os blocos projetados em nível de transistor. O circuito integrado é proposto para ser aplicado em identificação por rádio freqüência (RFID) na banda de UHF (Ultra High Frequency), usando multi-standard, e deve operar na faixa de 850 MHz a 1010 MHz. O sintetizador de freqüência foi projetado na tecnologia CMOS... / In this thesis, a frequency synthesizers phase locked loops (PLL) based with an architecture that uses a dual-path loop filter consisting of passive components and a digital integrator are proposed. The objective is to employ digital techniques to reduce the implementation cost and get loop filter design flexibility to enable the architecture to have a large tuning range operation and spurious reduction. The digital loop filter is based in an up/down counter where the phase difference is sampled to generate the clock of the counter. The techniques applied in the digital path are based in digitally controlled changes in the counter operation in predefined time points. These techniques provide PLL settling time reductions whiling the stability issues are solved. The stability study of the proposed dual path has been developed. First the linear system approximation for the PLL has been assumed and then employing digital control. The continuous and discrete time equations of architecture were derived in that study applied to stability design as well as to describe the architecture behavior. The top-down methodology has been applied to the integrated circuit design. In the beginning, the system level simulations are used for the techniques creation and then the behavioral models that were calibrated with transistor level blocks are simulated. The application of the circuit is proposed to Radio Frequency Identification (RFID) using UHF (Ultra High Frequency) band for multi-standards application and will operate in range of 850 MHz to 1010 MHz. The proposed frequency synthesizer has been designed in the AMS 0.35 μm CMOS technology with 2V power supply. A 300 μs of settling time and 140 Hz of resolution was obtained in simulations. The proposed frequency synthesizer have low complexity and shown a reference noise suppression about 45.6 dB better than the conventional architecture

Osciladores Controlados por Tensão (VCOs - Voltage-Controlled Oscillators) são circuitos de grande importância em sistemas de comunicação por radiofrequência atuais. Muitos trabalhos de pesquisa recentes têm focado no desenvolvimento de VCOs para aplicações em uma faixa muito grande de frequências (isto é, suportando amplo tunning range). O desenvolvimento de VCOs com uma ampla faixa de sintonia tem motivação na abertura de bandas de frequência, que até pouco tempo estavam licenciadas apenas para usos específicos, porém agora estão também abertas para a utilização de sistemas de rádios cognitivos. A ideia é que o rádio cognitivo tenha recursos para detectar se um canal (ou faixa de frequência) está sendo usado e, em caso de o canal não estar sendo usado, o rádio cognitivo deve se reconfigurar para operar nesse canal. Desse modo, os rádios cognitivos devem possuir um alto grau de reconfigurabilidade, de forma que possam operar em uma faixa muito ampla de frequências. Esse requisito exige o uso de de VCOs com um amplo tunning range. Este trabalho apresenta um projeto completo de um LC-VCO com uma larga faixa de frequência de operação (widedand). Um amplo tunning range foi obtido a partir do chaveamento (ou programação) do valor da capacitância total do tanque-LC do VCO, gerando assim várias sub-bandas de frequência. O ganho do VCO (KVCO) manteve-se com pequenas variações para todas as subbandas de frequência, com um valor médio de 88.6MHz, sendo 112MHz e 80MHz os valores máximo e mínimo, respectivamente. O ruído de fase variou de -118.4dBc/Hz a -107.4dBc/Hz para as portadores em 3.7GHz e 8.1GHz, respectivamente, enquanto que a potência dissipada do circuito LC-VCO variou de 1.8mW a 5.6mW para todo o tunning range. Para a figura de mérito power-frequency-tunning-normalized (FOMPFTN), os valores obtidos foram na faixa 3.1dB e 11.2dB, comparáveis com a maioria dos trabalhos publicados na área. / Voltage-Controlled Oscillators (VCOs) are very important circuits in current radio frequency communication systems. Much research has been focused recently on developing wideband VCOs in CMOS. The motivation on wideband VCOs is based on the opening of frequency bands, which until recently were licensed for specific uses, for use by cognitive radio systems. The idea is that cognitive radio must have the ability to detect whether a channel (or frequency band) is being used and if the channel is not being used, the cognitive radio must reconfigure itself to operate on that channel. Thus, cognitive radios should possess a high degree of reconfigurability, so that they can operate in a very wide frequency range. This requires the use of VCOs with a wide tunning range. This work presents a complete design of a LC-VCO with a wide operating frequency range (widedand). A wide tunning range has been obtained from the switching (or programming) the value of the total capacitance of the LC-tank of the VCO, thereby generating multiple frequency sub-bands. The VCO gain (KVCO) was maintained with small variations for all frequency sub-bands, with an average value of 88.6MHz, with 80MHz and 112MHz for the minimum and maximum values, respectively. The phase noise ranged from -118.4dBc/Hz to -107.4dBc/Hz for carriers at 3.7GHz and 8.1GHz, respectively, while the power dissipated in the LC-VCO circuit ranged from 1.8mW to 5.6mW for all tunning range. For the figure of merit power-frequency-tuning-normalized (FOMPFTN), the results were in the 3.1dB to 11.2dB range, comparable to most recently published works.

Made available in DSpace on 2014-06-11T19:24:46Z (GMT). No. of bitstreams: 0 Previous issue date: 2004-03-30Bitstream added on 2014-06-13T18:21:14Z : No. of bitstreams: 1 oliveira_vjs_me_ilha.pdf: 825294 bytes, checksum: 1231181cf2748d4fec35e435930c317b (MD5) / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / Nesse trabalho, propõe-se o projeto e implementação de um sintetizador de sinais ortogonais utilizando técnicas de circuito integrado e processo CMOS. O circuito do sintetizador baseia-se em um modelo matemático que utiliza multiplicadores e integradores analógicos, para geração de bases de funções ortogonais, tais como os polinômios de Legendre, as funções de base coseno e seno, a smoothed-cosine basis e os polinômios de Hermite. Funções ortogonais são bastante empregadas em processamento de sinais, e a implementação deste método matemático é capaz de gerar vários tipos de funções em um mesmo circuito integrado. O projeto proposto utiliza blocos analógicos funcionais para implementar o sintetizador. Os blocos que compõem o sintetizador foram projetados utilizando circuitos diferenciais, processamento em modo de corrente e técnicas de low-voltage. Algumas topologias utilizadas estão descritas na literatura, sendo que algumas foram adaptadas e mesmo modificadas, como no caso do multiplicador de corrente. Outras tiveram que ser propostas. As simulações e os resultados experimentais mostraram que o sintetizador é capaz de gerar funções ortogonais com amplitude e distorções satisfatórias. O sintetizador pode ser alimentado em 3V, tal qual foi projetado, tem faixa de entrada de ±20 μA e apresenta DHT (distorção harmônica total) inferior a 4% no quinto e último estágio em cascata. / In this work, a design and implementation of a synthesizer of orthogonal signals using CMOS technology and design technique for integrated circuits is proposed. The synthesizer circuit used analog multipliers and integrators for produce orthogonal functions such as Legendre polynomials, cosine and sine basis of functions, smoothed-cosine basis and Hermite polynomials. Orthogonal functions can be employed in signal processing and the implementation proposed can generate several kinds of functions in the same integrated circuit. In the synthesizer design building blocks was employed. The synthesizer’s blocks were design using differential circuits, low-voltage and current-mode techniques. Some topologies from papers were adapted or modified, as in the case of the current multiplier. Other topologies had to be proposed. The simulation and experimental results have shown that the synthesizer is able to produce orthogonal functions with satisfactory quality in distortions and amplitude. The synthesizer has a 3V supply voltage, a input current range of ±20 μA and it presents less than 4% of THD (Total Harmonic Distortion) in the last output in cascade.

Osciladores Controlados por Tensão (VCOs - Voltage-Controlled Oscillators) são circuitos de grande importância em sistemas de comunicação por radiofrequência atuais. Muitos trabalhos de pesquisa recentes têm focado no desenvolvimento de VCOs para aplicações em uma faixa muito grande de frequências (isto é, suportando amplo tunning range). O desenvolvimento de VCOs com uma ampla faixa de sintonia tem motivação na abertura de bandas de frequência, que até pouco tempo estavam licenciadas apenas para usos específicos, porém agora estão também abertas para a utilização de sistemas de rádios cognitivos. A ideia é que o rádio cognitivo tenha recursos para detectar se um canal (ou faixa de frequência) está sendo usado e, em caso de o canal não estar sendo usado, o rádio cognitivo deve se reconfigurar para operar nesse canal. Desse modo, os rádios cognitivos devem possuir um alto grau de reconfigurabilidade, de forma que possam operar em uma faixa muito ampla de frequências. Esse requisito exige o uso de de VCOs com um amplo tunning range. Este trabalho apresenta um projeto completo de um LC-VCO com uma larga faixa de frequência de operação (widedand). Um amplo tunning range foi obtido a partir do chaveamento (ou programação) do valor da capacitância total do tanque-LC do VCO, gerando assim várias sub-bandas de frequência. O ganho do VCO (KVCO) manteve-se com pequenas variações para todas as subbandas de frequência, com um valor médio de 88.6MHz, sendo 112MHz e 80MHz os valores máximo e mínimo, respectivamente. O ruído de fase variou de -118.4dBc/Hz a -107.4dBc/Hz para as portadores em 3.7GHz e 8.1GHz, respectivamente, enquanto que a potência dissipada do circuito LC-VCO variou de 1.8mW a 5.6mW para todo o tunning range. Para a figura de mérito power-frequency-tunning-normalized (FOMPFTN), os valores obtidos foram na faixa 3.1dB e 11.2dB, comparáveis com a maioria dos trabalhos publicados na área. / Voltage-Controlled Oscillators (VCOs) are very important circuits in current radio frequency communication systems. Much research has been focused recently on developing wideband VCOs in CMOS. The motivation on wideband VCOs is based on the opening of frequency bands, which until recently were licensed for specific uses, for use by cognitive radio systems. The idea is that cognitive radio must have the ability to detect whether a channel (or frequency band) is being used and if the channel is not being used, the cognitive radio must reconfigure itself to operate on that channel. Thus, cognitive radios should possess a high degree of reconfigurability, so that they can operate in a very wide frequency range. This requires the use of VCOs with a wide tunning range. This work presents a complete design of a LC-VCO with a wide operating frequency range (widedand). A wide tunning range has been obtained from the switching (or programming) the value of the total capacitance of the LC-tank of the VCO, thereby generating multiple frequency sub-bands. The VCO gain (KVCO) was maintained with small variations for all frequency sub-bands, with an average value of 88.6MHz, with 80MHz and 112MHz for the minimum and maximum values, respectively. The phase noise ranged from -118.4dBc/Hz to -107.4dBc/Hz for carriers at 3.7GHz and 8.1GHz, respectively, while the power dissipated in the LC-VCO circuit ranged from 1.8mW to 5.6mW for all tunning range. For the figure of merit power-frequency-tuning-normalized (FOMPFTN), the results were in the 3.1dB to 11.2dB range, comparable to most recently published works.

L'objectif de cette thèse est d'étudier la faisabilité et les performances d'un synthétiseur de fréquences agile pour les transmissions multi-bandes multi-utilisateurs destinées aux systèmes de transmission fournissant un très grand débit tout en répondant aux exigences de faible consommation et d'intégration facile. Dans ce contexte, les solutions classiques de synthétiseur de fréquences ne sont pas applicables et il est nécessaire de développer de nouvelles structures qui génèrent les fréquences centrales en permanence. La commutation d'une fréquence à l'autre peut se faire alors très rapidement par simple modification de la configuration des multiplexeurs. Dans la première partie de ce travail nous nous consacrons à l'analyse d'une telle architecture à fort potentiel. Une partie de sa structure est réalisée en technologie conventionnelle BiCMOS afin de valider sa faisabilité ainsi que le fonctionnement des solutions schématiques développées. Grâce à la caractérisation des composants, une analyse de la structure complète est réalisée. Cependant, la complexité de la structure du synthétiseur proposé fait de sorte qu'il existe pour les fréquences générées un grand nombre de fréquences parasites qui induisent des interférences entre utilisateurs. La réduction de leurs effets sur la transmission est l'objet de la deuxième partie de notre travail. Deux solutions numériques de réduction des interférences sont proposées. Elles permettent de rendre le design de la partie analogique moins contraignant en allégeant le cahier des charges et nous ont ainsi permis de simplifier l'architecture du synthétiseur. / The aim of this thesis is to study the feasibility and the performances of a fast switching frequency synthesizer designed for high debit multi-band multi-user transmission and used in transmission systems requiring a low consumption and an easy IC integration. In this context, the use of the classical synthesizer structures does not apply and there is a need to develop new architectures capable of generating all the frequencies permanently. Thus, the switching between frequencies can be easily done by changing multiplexors' state.In the first part of this study, we focus on the analysis of such high potential OL architecture. The main part of the proposed structure is implemented in a conventional BiCMOS technology in order to validate its feasibility and the operation of the developed blocks. Thanks to the measurements of the OL components, a complete analysis of the synthesizer is made. However, the complexity of the architecture of the proposed synthesizer induces the generation of large number of parasitic frequencies, creating interferences between the active users. The reduction of their effect on the transmission is the subject of the second part of the manuscript. Two digital methods are proposed to reduce the interferences. Lowering the requirements on the analog part, they allow a simplified design. This property was used to reduce the complexity of the frequency synthesizer.

Un des grands enjeux actuels dans le domaine de la synthèse sonore consiste à proposer des stratégies de contrôle intuitif accessible à un grand nombre d'utilisateurs. Parmi les attributs fondamentaux du son, ceux qui sont liés au mouvement restent encore largement inexplorés et, de fait, difficilement contrôlables. Cette thèse s'intéresse à la perception des caractéristiques des sons monophoniques qui varient au cours du temps et qui sont à l'origine de la notion perceptive de ''mouvement évoqué'', ainsi qu'à la mise au point d'une stratégie de contrôle intuitif. Pour cela, nous avons mis au point une série d'expériences basées sur des protocoles de catégorisation et d'évaluation sonore par le biais de dessins via une interface graphique paramétrée. Afin de favoriser l'écoute sur les attributs intrinsèques du son, nous avons privilégié l'utilisation de sons dits ''abstraits" pour lesquels la source physique n'est pas facilement identifiable. Les résultats de ces expériences nous ont permis de définir un ensemble de dimensions perceptives importantes pour la caractérisation du mouvement évoqué, telles que la forme de la trajectoire, l'orientation ou bien la taille. Ces dimensions ont été par la suite validées grâce à un protocole expérimental exploitant la synthèse sonore. Le travail réalisé ouvre de nombreuses perspectives, notamment dans les domaines de la réalité virtuelle, du ''design sonore'' et de la musique. La méthodologie proposée vise à être suffisamment générale pour être transposable à l'étude d'autres attributs perceptifs liés au ressenti (émotions, qualité ...). / An important issue within sound synthesis, is to propose intuitive control strategies that are accessible to most potential users. Among fundamental sound attributes, the ones linked to motion are almost unexplored and thus cannot be easily controlled. This Ph.D. concerns the perception of time varying characteristics of monophonic sounds, which leads to the notion of motion evoked by sounds and the conception of intuitive control strategies for such motions. To address those questions, we developed several experiments based on categorization tasks and evaluations through drawings using a parameterized graphical interface. In order to help listeners focus on intrinsic sound attributes, we mainly used so-called ''abstract'' sounds for which the physical sound source is not easily recognizable. The results of those experiments enabled us to identify attributes that characterize evoked motion from a perceptual point of view, such as shape, direction and size of the trajectory. Those dimensions were further validated in an experimental protocol using sound synthesis. This work opens many perspectives, in particular within virtual reality, sound design and musical domains. The proposed methodology aimed at being sufficiently generic to be transposed to other perceptual attributes such as emotions and quality.

L’ambiance sonore dans l’habitacle tend à être mieux maitrisée grâce au progrès de l’isolation et à l’introduction de motorisations plus silencieuses. Dans le cas des véhicules thermiques, il est désormais possible de modifier ou d’ajouter des organes au moteur pour rendre sa sonorité plus expressive et améliorer ainsi la sensation d’accélération. La synthèse sonore permet de simuler ces différents réglages du moteur et d’étudier leur ressenti. Pour être simple et efficace, cette synthèse doit donc répondre aux attentes des concepteurs sonores, par exemple donner une sonorité sportive au moteur.Le bruit dans l’habitacle est un bruit complexe puisque son timbre varie avec la dynamique du véhicule. Sa description perceptive est connue, notamment par l’emploi d’onomatopées (« ON », « AN » et « REU »), mais la caractérisation des ces descripteurs demeurent incomplètes. Il est donc difficile de manipuler les paramètres du signal lors de la synthèse pour reproduire ces attributs perceptifs dans les sons créés. Notre but est de proposer une nouvelle synthèse pour établir ce lien manquant entre perception et signal.Une expérience en simulateur de conduite étudie tout d’abord le couplage entre le ressenti acoustique et le ressenti dynamique du véhicule. Ensuite, nous cherchons à établir le lien entre la perception du bruit moteur et la synthèse par l’utilisation d’imitations vocales du bruit moteur reprenant les onomatopées « ON » et « AN ». Une modélisation du bruit dans l’habitacle est ainsi construite en s’inspirant d’un modèle source/filtre, puis testée dans deux expériences en laboratoire acoustique. Le bruit moteur peut alors être « métaphoriquement » assimilé aux cordes vocales du véhicule tandis que les résonances de l’habitacle sont considérées comme le conduit vocal. / Automotive acoustics is living a new challenge due to the introduction of new power-trains. Therefore, interior car noise is being well designed, offering new perspectives in terms of sound ambiance. The interior car noise for combustion engine cars can be modified in order to increase the comfort by producing the impression of a quiet car, or oppositely, by rendering it more expressive, the driver feels better the dynamics of the car. Currently, these sensations are produced by modifying the engine components. But a very interesting alternative is presented by sound synthesis, which gives a new dimension to sound conception. Perceptive studies and synthesis control contributes to develop this technology. Engine noise is a complex noise with a timbre varying with car dynamics, and it can be described using perceptual descriptors, like onomatopoeia (“ON”, “AN” and “REU”). However, the use of perceptive descriptors does not allow finding the link between them and the physical characteristics of engine sound, therefore there they are difficult to integrate in sound synthesis. The goal of this research is to produce a sound synthesis based on human sound perception and car dynamics. A first study was developed on a driving simulator in order to describe the relation between acoustic perception and motion perception. The next step was to establish the link between perception and synthesis by vocal imitation based on onomatopoeia “ON” and “AN” reproducing engine noise. Based on these results, a subtractive synthesis of interior car noise was further built, inspired by a source/filter model. Last, but not least, we tested the impact of engine noise and car interior resonance in two experiments in acoustic laboratory. Engine noise can be metaphorically considered as the car “vocal chords”, while the resonance of the interior of the vehicle can be considered as its “vocal tract”.

The sound synthesizer is an electronic musical instrument that has become commonplace in audio production for music, film, television and video games. Despite its widespread use, creating new sounds on a synthesizer - referred to as synthesizer programming - is a complex task that can impede the creative process. The primary aim of this thesis is to support the development of techniques to assist synthesizer users to more easily achieve their creative goals. One of the main focuses is the development and evaluation of algorithms for inverse synthesis, a technique that involves the prediction of synthesizer parameters to match a target sound. Deep learning and evolutionary programming techniques are compared on a baseline FM synthesis problem and a novel hybrid approach is presented that produces high quality results in less than half the computation time of a state-of-the-art genetic algorithm. Another focus is the development of intuitive user interfaces that encourage novice users to engage with synthesizers and learn the relationship between synthesizer parameters and the associated auditory result. To this end, a novel interface (Synth Explorer) is introduced that uses a visual representation of synthesizer sounds on a two-dimensional layout. An additional focus of this thesis is to support further research in automatic synthesizer programming. An open-source library (SpiegeLib) has been developed to support reproducibility, sharing, and evaluation of techniques for inverse synthesis. Additionally, a large-scale dataset of one billion sounds paired with synthesizer parameters (synth1B1) and a GPU-enabled modular synthesizer (torchsynth) are also introduced to support further exploration of the complex relationship between synthesizer parameters and auditory results. / Graduate

Le regard de la société sur les avancées technologiques a été la cible de multiples remises en question. Des premiers synthétiseurs aux allures monstrueuses et presque grotesques aux instruments virtuels, les compositeurs de musique de film ont souvent port. un regard curieux sur cet instrument énigmatique, aussi singulier que versatile. Un synthétiseur est un appareil musical électronique ou un logiciel utilisé pour générer et combiner des signaux audio par des processus analogiques ou numériques . (traduit de Devine, 2013). Par un processus d’expérimentation, certains compositeurs ont relevé le défi musical d’utiliser le synthétiseur dans la musique de film. Avec le temps, ils ont trouvé en lui un outil irremplaçable. / Society's view of technological advances has been the target of many questionings. From the first monstrous and almost grotesque looking synthesizers to virtual instruments, composers of film music have often taken a curious look at this enigmatic instrument, as singular as it is versatile. A synthesizer is an "electronic musical device or software used to generate and combine audio signals through analog or digital processes" (translated from Devine, 2013). Through a process of experimentation, some composers have taken on the musical challenge of using the synthesizer in film music. Over time, they have found in it an irreplaceable tool.

This report details the process of developing libraries to handle data management on two types of non-volatile memory interfacing with an STM32 microprocessor. This was done at the request of Suonobouno AB, who are developing digital and analog synthesizers. This work included the development of generic open-ended memory management functions, as well as functions for handling the specific data structures used in the company's instruments, the later building on the former. In the end, a library was developed that can perform the reading and writing of data to the storage mediums through a clean interface devoid of unnecessary detail. The solution is highly modular, with a clear strategy for managing coupling, to allow the library to grow and develop with further integration. / I denna rapport beskrivs arbetet med att utveckla bibliotek för hantering av externa lagringsenheter som används av en STM32 mikroprocessor. Detta utfördes på uppdrag av Suonobouno AB som utvecklar digitala och analoga syntar. Detta arbete innefattade framtagandet av generiska funktioner för minneshantering, så väl som funktioner anpassade för de specifika datastrukturer som förekommer i det system som är den tilltänkta applikationen. I slutändan åstadkoms ett bibliotek som kan utföra den nödvändiga läsningen och lagringen av data till de två lagringsmedia som omfattades av projektet. Detta kan kontrolleras genom ett rent interface som inte kräver kännedom om tekniska detaljer bakom lagringsteknikerna. Biblioteket är modulärt, med en tydlig strategi för att hantera beroenden, så att det kan växa och utvecklas med vidare integrering.