Papers by Stéphane Lesoinne
NOISE-CON ... proceedings, Nov 29, 2023
In oshore wind farms, the installation of a wind turbine is commonly done by driving a pile into ... more In oshore wind farms, the installation of a wind turbine is commonly done by driving a pile into the seabed, generating high energetic impulsive noise with possible negative impact on the marine fauna. The assessment of the environmental impact of pile driving has becomed an important challenge. In conjunction with numerical models of the sound source (pile driven by hammer), numerical underwater sound propagation models can be used to predict the sound pressure levels evolution over the distance from the pile. However, those models depend on environmental input data which are marred by uncertainties that can aect the simulation results. In this paper, the impact of those uncertainties on far-eld propagation transmission losses is studied using an elastic parabolic equation (RAMS) on a test case in the English Channel. The sensitivity of the model is given for the estimated maximum boundaries on the data uncertainties.
Journal of the Acoustical Society of America, May 1, 2008
Randomized ray tracing in room acoustics can be used to compute echograms, but the results comput... more Randomized ray tracing in room acoustics can be used to compute echograms, but the results computed at the receptor are affected by statistical errors. To decrease these statistical errors, the usual solution is to increase the number of rays, but this also increases the computation time. Another solution is to increase the receptor size, but this lowers the spatial resolution of the computed sound field, which is particularly important for the early part of the echogram. In order to decrease the computation time and keep a sufficient resolution, the method proposed in this paper is based on a progressive modification of the receptor size (spherical type) as long as the ray's travel grows. At the same time, the number of rays into the room can be decreased as the sound ray lengths increase while keeping the statistical errors more or less constant. The consequence is that the sound field spatial resolution evolves from "precise" at the beginning of the echogram to "rough" at its end. The first applications of this new method show a significant gain of CPU time.
Journal of The Audio Engineering Society, May 1, 2005
Strong insulation in modern building and housing requires efficient ventilation systems. A popula... more Strong insulation in modern building and housing requires efficient ventilation systems. A popular solution is a decentralized setup with one ventilation unit per room. The noise emitted by a ventilation unit then has a major impact on the room comfort. In order to reduce the heat transfer to the outside, a ventilation unit is typically designed as a double flux-system with a heat exchanger. This exchanger has a noticeable impact on the acoustic behavior of the ventilation unit. It is therefore of interest to study its effect through numerical simulations. The numerical modelling of a realistic heat exchanger is presented in this paper. The exchanger is placed inside its ventilation casing and modelled using a double equivalent fluid homogenization. Unknown homogenization properties are retrieved in two steps. Acoustic transfer functions are first measured experimentally in order to characterize the propagation paths through the heat exchanger. In a second step, an optimization loop is computed with the numerical model of the heat exchanger. This allows to determine the homogenization properties fitting the measured transfer functions for each frequency. Finally the results for the characterized heat exchanger homogenization model are compared against measurement.
HAL (Le Centre pour la Communication Scientifique Directe), Apr 12, 2010
Dans un algorithme classique de tir de rayons sonores, la recherche de l'intersection entre un se... more Dans un algorithme classique de tir de rayons sonores, la recherche de l'intersection entre un segment de rayon sonore et la prochaine surface est effectuée en testant l'intersection de ce rayon avec toutes les surfaces présentes. Cependant, dans une salle quelconque, la probabilité d'intersection d'un rayon avec le modèle géométrique n'est pas uniformément répartie entre les surfaces. Pour chaque surface « i » origine d'un segment de rayon, l'idée est d'obtenir la liste des surfaces « j » triée par ordre de probabilité décroissante d'intersection. Cette probabilité d'intersection peut-être estimée à partir d'un petit sous-ensemble du nombre total de rayons sonores tirés par la source acoustique. Lors d'une recherche, la liste des surfaces triées peut être parcourue selon différentes stratégies :-Assigner la première surface trouvée dans la liste comme surface destination, au prix d'une imprécision sur les trajets des rayons. En fonction du soin apporté à la modélisation de la salle, cette solution peut entraîner la perte de nombreux rayons.-Interrompre la recherche à la première intersection trouvée et chercher s'il existe une intersection avec les surfaces intermédiaires, surfaces situées entre la surface origine du segment et la surface destination trouvée. La liste des surfaces intermédiaires est obtenue avant le tir de rayons au moyen d'enveloppes convexes. Cette modification permet de ne perdre aucun rayon et de respecter le trajet initial des rayons.
Strong insulation in modern buildings and housing requires efficient ventilation systems. A popul... more Strong insulation in modern buildings and housing requires efficient ventilation systems. A popular solution is a decentralized system with one ventilation unit per room, but the noise emitted by such a unit then has a major impact on the comfort. In this research, it is intended to equip an existing ventilation unit with an active noise control (ANC) system to reduce its low-frequency noise emissions. The ANC system will be hosted in an additional duct deporting the air inlet aperture approximately one meter away from its original position. The one-channel feedforward ANC system has a reference microphone located between the fan and the anti-noise loudspeaker and an error microphone at the duct's end. The transfer functions have been computed by a 3D FEM solver, between the fan and several reference and error microphone locations. Several combinations of reference/error microphone positions were then tested regarding the possible active attenuation of a white noise emitted by the primary source. Optimal control by the ANC controller was assumed, the optimal filter being computed with the help of the corresponding transfer functions. Theoretical attenuations were obtained and optimal positions were finally defined for the two microphones of the ANC system. The first results are finally discussed in this paper and raise some questions regarding the influence of the noise spectrum on the attenuations obtained.
Auralization in room acoustics is created by the convolution of an anechoic audio signal with the... more Auralization in room acoustics is created by the convolution of an anechoic audio signal with the RIR (room impulse response), either computed or measured at the receiver location. When headsets are used for reproduction, the same auralized signal is often sent to both ears, for simplicity. In this paper, we introduce the concept of "directional room impulse responses" and their computation with a sound ray program is reported. These responses are used in the convolution process in combination with HRTFs (head related transfer functions) to simulate not only the room reverberation, but also to accurately locate the direct sound, image sources and, in general, all contributions to the receiver. It is assumed here that headsets are used for reproduction, but it is obvious that directional RIRs are not limited to this application. Convolving each sound ray reaching the receiver with the corresponding HRTF is not practical and does not lead to correct auralization. Therefore, the technique we use is based upon the separation of the directional RIR in two contributions: a first one including the direct sound and the specular reflections up to a given order, the second contribution including all the other contributions (including diffuse reflections). Computational problems include the separate treatment of specular and diffuse reflections and the compromise between computing time, the number and the accuracy of directional RIRs.
NOISE-CON ... proceedings, Feb 1, 2023
In a ventilation duct network, two types of silencers can be used: passive (medium and high frequ... more In a ventilation duct network, two types of silencers can be used: passive (medium and high frequencies) and active (low frequencies). Active silencers have a loudspeaker (control loudspeaker) that generates an anti-noise. Usually, this system is placed close to the fan and far from the duct air opening. However, when the system is placed close to an air opening, the control speaker can not only be used to reduce the fan noise, but also as a background speaker to play music, information, alerts (the target signal) ... This concept has been implemented in an active noise control (ANC) system based on the adaptive feedforward normalized FxLMS algorithm which uses a reference and an error microphone to pick up respectively the noise to attenuate and the residual noise after control to be minimized. The algorithm is modified so that the target signal does not contaminate the reference nor the error signals. The new algorithm is tested in real time in a short duct with different type of noises and target signals.
Journal of The Audio Engineering Society, 2005
The Journal of the Acoustical Society of America, 2008
Randomized ray tracing in room acoustics can be used to compute echograms, but the results comput... more Randomized ray tracing in room acoustics can be used to compute echograms, but the results computed at the receptor are affected by statistical errors. To decrease these statistical errors, the usual solution is to increase the number of rays, but this also increases the computation time. Another solution is to increase the receptor size, but this lowers the spatial resolution of the computed sound field, which is particularly important for the early part of the echogram. In order to decrease the computation time and keep a sufficient resolution, the method proposed in this paper is based on a progressive modification of the receptor size (spherical type) as long as the ray's travel grows. At the same time, the number of rays into the room can be decreased as the sound ray lengths increase while keeping the statistical errors more or less constant. The consequence is that the sound field spatial resolution evolves from "precise" at the beginning of the echogram to "rough" at its end. The first applications of this new method show a significant gain of CPU time.
Auralization in room acoustics is created by the convolution of an anechoic audio signal with the... more Auralization in room acoustics is created by the convolution of an anechoic audio signal with the RIR (room impulse response), either computed or measured at the receiver location. When headsets are used for reproduction, the same auralized signal is often sent to both ears, for simplicity. In this paper, we introduce the concept of "directional room impulse responses" and their computation with a sound ray program is reported. These responses are used in the convolution process in combination with HRTFs (head related transfer functions) to simulate not only the room reverberation, but also to accurately locate the direct sound, image sources and, in general, all contributions to the receiver. It is assumed here that headsets are used for reproduction, but it is obvious that directional RIRs are not limited to this application. Convolving each sound ray reaching the receiver with the corresponding HRTF is not practical and does not lead to correct auralization. Therefore, ...
In o shore wind farms, the installation of a wind turbine is commonly done by driving a pile into... more In o shore wind farms, the installation of a wind turbine is commonly done by driving a pile into the seabed, generating high energetic impulsive noise with possible negative impact on the marine fauna. The assessment of the environmental impact of pile driving has becomed an important challenge. In conjunction with numerical models of the sound source (pile driven by hammer), numerical underwater sound propagation models can be used to predict the sound pressure levels evolution over the distance from the pile. However, those models depend on environmental input data which are marred by uncertainties that can a ect the simulation results. In this paper, the impact of those uncertainties on fareld propagation transmission losses is studied using an elastic parabolic equation (RAMS) on a test case in the English Channel. The sensitivity of the model is given for the estimated maximum boundaries on the data uncertainties.
Dans un algorithme classique de tir de rayons sonores, la recherche de l’intersection entre un se... more Dans un algorithme classique de tir de rayons sonores, la recherche de l’intersection entre un segment de rayon sonore et la prochaine surface est effectuee en testant l’intersection de ce rayon avec toutes les surfaces presentes. Cependant, dans une salle quelconque, la probabilite d’intersection d’un rayon avec le modele geometrique n’est pas uniformement repartie entre les surfaces. Pour chaque surface « i » origine d’un segment de rayon, l’idee est d’obtenir la liste des surfaces « j » triee par ordre de probabilite decroissante d’intersection. Cette probabilite d’intersection peut-etre estimee a partir d’un petit sous-ensemble du nombre total de rayons sonores tires par la source acoustique. Lors d’une recherche, la liste des surfaces triees peut etre parcourue selon differentes strategies : Assigner la premiere surface trouvee dans la liste comme surface destination, au prix d’une imprecision sur les trajets des rayons. En fonction du soin apporte a la modelisation de la salle...
Strong insulation in modern buildings and housing requires efficient ventilation systems. A popul... more Strong insulation in modern buildings and housing requires efficient ventilation systems. A popular solution is a decentralized system with one ventilation unit per room, but the noise emitted by such a unit then has a major impact on the comfort. In this research, it is intended to equip an existing ventilation unit with an active noise control (ANC) system to reduce its low-frequency noise emissions. The ANC system will be hosted in an additional duct deporting the air inlet aperture approximately one meter away from its original position. The one-channel feedforward ANC system has a reference microphone located between the fan and the anti-noise loudspeaker and an error microphone at the duct’s end. The transfer functions have been computed by a 3D FEM solver, between the fan and several reference and error microphone locations. Several combinations of reference/error microphone positions were then tested regarding the possible active attenuation of a white noise emitted by the p...
Uploads
Papers by Stéphane Lesoinne