143 research outputs found
A survey of sound source localization methods in wireless acoustic sensor networks
No full textWireless acoustic sensor networks (WASNs) are formed by a distributed group of acoustic-sensing devices featuring audio playing and recording capabilities. Current mobile computing platforms offer great possibilities for the design of audio-related applications involving acoustic-sensing nodes. In this context, acoustic source localization is one of the application domains that have attracted the most attention of the research community along the last decades. In general terms, the localization of acoustic sources can be achieved by studying energy and temporal and/or directional features from the incoming sound at different microphones and using a suitable model that relates those features with the spatial location of the source (or sources) of interest. This paper reviews common approaches for source localization in WASNs that are focused on different types of acoustic features, namely, the energy of the incoming signals, their time of arrival (TOA) or time difference of arrival (TDOA), the direction of arrival (DOA), and the steered response power (SRP) resulting from combining multiple microphone signals. Additionally, we discuss methods not only aimed at localizing acoustic sources but also designed to locate the nodes themselves in the network. Finally, we discuss current challenges and frontiers in this field
Zero-shot anomalous sound detection in domestic environments using large-scale pretrained audio pattern recognition models
No full textAnomalous sound detection is central to audio-based surveillance and monitoring. In a domestic environment, however, the classes of sounds to be considered anomalous are situation-dependent and cannot be determined in advance. At the same time, it is not feasible to expect a demanding labeling effort from the end user. To address these problems, we present a novel zero-shot method relying on an auxiliary large-scale pretrained audio neural network in support of an unsupervised anomaly detector. The auxiliary module is tasked to generate a fingerprint for each sound occasionally registered by the user. These fingerprints are then compared with those extracted from the input audio stream, and the resulting similarity score is used to increase or reduce the sensitivity of the base detector. Experimental results on synthetic data show that the proposed method substantially improves upon the unsupervised base detector and is capable of outperforming existing few-shot learning systems developed for machine condition monitoring without involving additional training
Acoustic source localization in the spherical harmonics domain exploiting low-rank approximations
No full textAcoustic signal processing in the spherical harmonics domain (SHD) is an
active research area that exploits the signals acquired by higher order
microphone arrays. A very important task is that concerning the localization of
active sound sources. In this paper, we propose a simple yet effective method
to localize prominent acoustic sources in adverse acoustic scenarios. By using
a proper normalization and arrangement of the estimated spherical harmonic
coefficients, we exploit low-rank approximations to estimate the far field
modal directional pattern of the dominant source at each time-frame. The
experiments confirm the validity of the proposed approach, with superior
performance compared to other recent SHD-based approaches.Comment: To appear in ICASSP 202
Deep-Learning-Based Radio Map Reconstruction for V2X Communications
No full textRadio environment map (REM) reconstruction based on large-scale channel measurements is a promising technology for future mobility services involving vehicle-to-everything (V2X) communications. REMs provide contextual information which can be exploited to reduce V2X communication latency and control signaling, for instance, through a fast access to channel state information. However, the accuracy of radio mapping techniques is limited by the availability of measurements, which require for collection significant signaling overhead. Moreover, mobility scenarios impose strict latency constraints that render fast channel acquisition a challenging problem. This paper presents a low-complexity deep-learning-based approach based on long-short term memory (LSTM) cells for REM reconstruction on roads, addressed as a data-filling problem. To improve model generalization, the network is trained on a virtually infinite dataset generated according to a 3GPP-compliant freeway scenario, considering different correlation properties and missing point configurations. The results show that the proposed approach provides a performance closer to the theoretical lower bound than the classical Ordinary Kriging spatial interpolation method, without increasing the complexity order. Experiments performed in realistic scenarios using a 3D city model confirm the generalization capability of the proposed solution
Parametric head-related transfer function modeling and interpolation for cost-efficient binaural sound applications
No full textParametric methods for modeling the perceptually relevant features of head-related transfer functions (HRTFs) are very important for the development of low-cost immersive sound applications. This letter describes an efficient method based on a low-order infinite impulse response filter implemented by a chain of second order sections of conventional shelving and peak audio filters. The parameters (central frequency, gain, and quality factor) are numerically adjusted by iteratively fitting the frequency response of the filter to the desired HRTF. Besides allowing for low-order binaural models, the proposed approach provides an efficient way to synthesize HRTFs for non-measured angles by applying a simple interpolation between the parameters from neighboring responses. Additionally, the HRTF database size is significantly reduced. (C) 2013 Acoustical Society of America.The Spanish Ministry of Economy and Competitiveness and FEDER supported this work under the project TEC2012-37945-C02-01/02. Part of this work was also funded by Generalitat Valenciana Grant BEST2010.Ramos Peinado, G.; Cobos Serrano, M. (2013). Parametric head-related transfer function modeling and interpolation for cost-efficient binaural sound applications. Journal of the Acoustical Society of America. 134(3):1735-1738. https://doi.org/10.1121/1.4817881S17351738134
Aisopou Mythoi/Fabulae Aesopicae Graecae Quae Maximo Planudi Tribuuntur
No full textThe full title reads this way: Aisopou mythoi Fabulae Aesopicae Graecae quae Maximo Planudi tribuuntur ad veterum librorum fidem emendatas Ioannis Hudsonis suisque adnotationibus illustratas atque indice verborum locupletissimo instructas edidit Io. Michael Heusinger curavit et praefatus est Christ. Adolph. Klotzius. The book is easy to find online; it seems to be the only book that comes up if one Googles Wittekindt, although this edition may be a reprint of a 1770 edition also done by Wittekindt. The book has three parts: a lengthy unpaginated introduction; 120 pages of some 149 fables and their variants; and a Greek/Latin dictionary of all the vocables that occur in Aesop's fables. This third section has a two-page addition cataloguing items explained in the notes. The whole comes, online sources say, to 288 pages. The introductory section includes a preface by Klotz; a catalogue of manuscripts by Heusinger, a preface to the reader by Hudson; and sixty-eight ancient testimonies to Aesop and the fables. For those confused by the shortened and Latinized names on the title-page, here are more complete and modern names: Christian Adolph Klotz edited this book, following on the work of editor and commentator John Hudson with Johann Michael Heusinger's index verborum. The publisher is Johann Georg Ernst Wittekindt.This is a hardbound book (hard cover)Language note: GreekJoannis Hudsonis, Io. Michael Heusinger, Christ. Adolph. Klotziu
Time Difference of Arrival Estimation from Frequency-Sliding Generalized Cross-Correlations Using Convolutional Neural Networks
No full textThe interest in deep learning methods for solving traditional signal processing tasks has been steadily growing in the last years. Time delay estimation (TDE) in adverse scenarios is a challenging problem, where classical approaches based on generalized cross-correlations (GCCs) have been widely used for decades. Recently, the frequency-sliding GCC (FS-GCC) was proposed as a novel technique for TDE based on a sub-band analysis of the cross-power spectrum phase, providing a structured two-dimensional representation of the time delay information contained across different frequency bands. Inspired by deep-learning-based image denoising solutions, we propose in this paper the use of convolutional neural networks (CNNs) to learn the time-delay patterns contained in FS-GCCs extracted in adverse acoustic conditions. Our experiments confirm that the proposed approach provides excellent TDE performance while being able to generalize to different room and sensor setups
A note on the modified and mean-based steered-response power functionals for source localization in noisy and reverberant environments
No full textCombinación de cuestionarios simples y gamificados utilizando gestores de participación en el aula: experiencia y percepción del alumnado
No full text[EN] The growing use of mobile devices has motivated the development of a wide range of applications to help manage the students’ participation in the classroom. Socrative allows the lecturer to use multiple-choice questionnaires in the classroom, either in a simple or a gamified mode (Space Race). In this paper, we describe our experience at using this tool to promote competitive learning, at both undergraduate and post-graduate levels. The student’s perception indicates that the use of the application helped at increasing engagement and motivation. However, relevant differences were found between both modes of use, underlining the importance of an adequate activity design.[ES] En los últimos años y gracias a la utilización masiva de dispositivos móviles,han proliferado múltiples aplicaciones para la gestion de la participación del alumnado en el aula. Específicamente, Socrative, permite ellanzamiento por el profesor de cuestionarios de opción múltiple en modosimple o en modo gamicado (carrera espacial), fomentando así el juegoy la competitividad. En este trabajo se describe la experiencia obtenidamediante la combinación de ambos modos en una asignatura de Grado yotra de Máster. Los resultados obtenidos as como la percepcion del alumnadoindican que los alumnos se ven altamente motivados por este tipo deherramientas, si bien existen diferencias importantes en ambos modos deutilizacion que conviene tener presentes en el diseño de la actividad.Roger, S.; Cobos, M.; Arevalillo-Herráez, M.; García-Pineda, M. (2017). Combinación de cuestionarios simples y gamificados utilizando gestores de participación en el aula: experiencia y percepción del alumnado. En In-Red 2017. III Congreso Nacional de innovación educativa y de docencia en red. Editorial Universitat Politècnica de València. 1128-1139. https://doi.org/10.4995/INRED2017.2017.6746OCS1128113
- …
