Bruit de fond | EPFL Graph Search

En traitement du signal, on appelle bruit de fond toute composante non désirée affectant la sortie d'un dispositif indépendamment du signal présent à son entrée. Le bruit de fond se décompose en bruit propre, que cause le dispositif lui-même, et en perturbations originaires de l'extérieur qu'il capte malencontreusement. Importance et caractérisation Au fur et à mesure que le signal se rapproche, puis s'enfonce en dessous du niveau du bruit de fond, la quantité d'informations qu'il peut transporter décroît , il devient plus difficile à détecter, et il finit par se dissoudre dans l'incertitude. Répartition spectrale du bruit Le bruit de fond perturbe d'autant plus un signal qu'il occupe la même bande de fréquences. Modèles mathématiques L'analyse spectrale d'un signal électrique débouche sur les notions de largeur de bande et de bande passante. L'étude d'un canal de communication s'effectue alors par des fonctions sinusoïdales. On peut aussi

De l'acoustique des églises en Suisse

Victor Desarnaulds

The existence of the fundamental conflict in churches between the desirable acoustic conditions for music and for speech intelligibility prompted this dissertation. It tackles the subject not only under the acoustical perspective but also in the historical, theological, musicological and architectural aspects. This research studies, initially, how during the past, the liturgy, the acoustical and organizational characteristics of churches supported one or the other of these attributes. The second chapter presents the objective analyses of many Reformed worships and Catholic masses which make it possible to show the conditions and requirements for the current Christian celebrations in Switzerland. The analysis of the church geometrical parameters (volume, height, etc.) and of the results of objective acoustical measurements (reverberation time, background noise and speech intelligibility) of approximately two hundred churches in Switzerland allows to statistically describe their characteristics and to objectively verify the assumptions made in the preceding chapter. Chapter 3 also studies the local or global range of conclusions by the comparison of the results obtained in Switzerland with those published for other countries (synthesis of about 400 churches). The study of the origin and the interaction between Liturgy and Acoustics passes in particular by the analysis of the history of sacred music but especially by the examination of the main methods implemented to improve the acoustics of churches. Chapter 4 presents the historical evolution and the acoustical effectiveness of two types of devices used in churches: acoustic pots and pulpits with their canopies. The position of the various participants does not influence only the listening conditions but also reveals the importance attached to their participation. Chapter 5 examines then the historical evolution and statistical analysis of the current conditions regarding the position of the organ, the spatial arrangements for the seated congregation in churches, as well as the effect of the occupation on reverberation time and speech intelligibility. Chapter 6 concludes with the analysis of the desirable acoustical conditions in current churches, particularly with regard to the architectural, organizational (sound system, assembly arrangement and organ location) and the acoustical parameters (reverberation time, background noise and speech intelligibility). This last analysis is based on a significant bibliographical synthesis and on the comparison, for the sampled churches, of the results for objective measurements with the subjective evaluations obtained by a parishioner's sample survey.

EPFL2002

Combinatorial algorithms for wireless information flow

Aurore Amaudruz, Christina Fragouli

A long-standing open question in information theory is to characterize the unicast capacity of a wireless relay network. The difficulty arises due to the complex signal interactions induced in the network, since the wireless channel inherently broadcasts the signals and there is interference among transmissions. Recently, Avestimehr, Diggavi and Tse proposed a linear binary deterministic model that takes into account the shared nature of wireless channels, focusing on the signal interactions rather than the background noise. They generalized the min-cut max-flow theorem for graphs to networks of deterministic channels and proved that the capacity can be achieved using information theoretical tools. They showed that the value of the minimum cut is in this case the minimum rank of all the binary adjacency matrices describing source-destination cuts. However, since there exists an exponential number of cuts, identifying the capacity through exhaustive search becomes infeasible. In this work, we develop a polynomial time algorithm that discovers the relay encoding strategy to achieve the min-cut value in binary linear deterministic (wireless) networks, for the case of a unicast connection. Our algorithm crucially uses a notion of linear independence between edges to calculate the capacity in polynomial time. Moreover, we can achieve the capacity by using very simple one-bit processing at the intermediate nodes, thereby constructively yielding finite length strategies that achieve the unicast capacity of the linear deterministic (wireless) relay network.

2009

Deep Learning Approaches for Auditory Perception in Robotics

Weipeng He

Auditory perception is an essential part of a robotic system in Human-Robot Interaction (HRI), and creating an artificial auditory perception system that is on par with human has been a long-standing goal for researchers. In fact, this is a challenging research topic, because in typical HRI scenarios the audio signal is often corrupted by the robot ego noise, other background noise and overlapping voices. The traditional approaches based on signal processing seek analytical solutions according to the physical law of sound propagation as well as assumptions about the signal, noise and environments. However, such approaches either assume over-simplified conditions, or create sophisticated models that do not generalize well in real situations. This thesis introduces an alternative methodology to auditory perception in robotics by using deep learning techniques. It includes a group of novel deep learning-based approaches addressing sound source localization, speech/non-speech classification, and speaker re-identification. The deep learning-based approaches rely on neural network models that learn directly from the data without making many assumptions. They are shown by experiments with real robots to outperform the traditional methods in complex environments, where there are multiple speakers, interfering noises and no a priori knowledge about the number of sources. In addition, this thesis addresses the issue of high cost of data collection which arises with learning-based approaches. Domain adaptation and data augmentation methods are proposed to exploit simulated data and weakly-labeled real data, so that the effort for data collection is minimized. Overall, this thesis suggests a practical and robust solution for auditory perception in robotics in the wild.

EPFL2021

De l'acoustique des églises en Suisse

Victor Desarnaulds

EPFL2002