Audio Processing

This paper presents an automated Public Address processing unit, using delay and magnitude response adjustment. The aim is to achieve a flat frequency response and delay adjustment between different physically-placed speakers at the measuring point, which is nowadays usually made manually by the sound technician. The adjustment is obtained using three signal processing operations to the audio signal: time delay adjustment, crossover filtering, and graphic equalization. The automation is in the calculation of different parameter sets: estimation of the time delay, the selection of a suitable crossover frequency, and calculation of the gains for a third-octave graphic equalizer. These automatic methods reduce time and effort in the calibration of line-array PA systems, since only three sine sweeps must be played through the sound system. Measurements have been conducted in an anechoic chamber using a 1:10 scale model of a line array system to verify the functioning of the automatic calibration and equalization methods.

This paper presents an automated Public Address processing unit, using delay and magnitude response adjustment. The aim is to achieve a flat frequency response and delay adjustment between different physically-placed speakers at the measuring point, which is nowadays usually made manually by the sound technician. The adjustment is obtained using three signal processing operations to the audio signal: time delay adjustment, crossover filtering, and graphic equalization. The automation is in the calculation of different parameter sets: estimation of the time delay, the selection of a suitable crossover frequency, and calculation of the gains for a third-octave graphic equalizer. These automatic methods reduce time and effort in the calibration of line-array PA systems, since only three sine sweeps must be played through the sound system. Measurements have been conducted in an anechoic chamber using a 1:10 scale model of a line array system to verify the functioning of the automatic ca...

Retrieving a music file from a large database is a non-trivial task. To support this task, many mechanisms have been developed over the years. However, indexing files remains one of the most popular mechanisms. Several algorithms allow feature extraction from audio signals. Usually, these features are used to describe music content. In this paper, we demonstrate that associations between content-based descriptors can be used as well. We have developed a processing chain which uses association rules mining to find significant relations between content-based descriptors of music files. The significant relations are used to index music files. Experiments conducted demonstrates that the proposed approach can yield interesting results especially with classical music.

Over the last few years, most of the tasks employing Deep Learning techniques for audio processing have achieved stateof-the-art results employing Conformer-based systems. However, when it comes to sound event detection (SED), it was scarcely used after it won the DCASE Challenge 2020 Task 4. In previous research, we found that Conformer-based systems achieved a higher performance in terms of sound events classification compared to other architectures frequently employed, such as Convolutional Recurrent Neural Networks (CRNNs). Given that the second scenario proposed for the Polyphonic Sound Detection Score (PSDS2) is focused on avoiding confusion between classes, in this paper we propose to optimize a Conformer-based system to maximize the performance on this scenario. For this purpose, we performed a hyperparameter tuning and incorporated recently proposed Frequency Dynamic Convolutions (FDY) to enhance its classification properties. Additionally, we employed our previously proposed multi-resolution approach not only to enhance the performance but also to gain a deeper understanding of the Conformer architecture for SED, analyzing its advantages and disadvantages, and finding possible solutions to them. Additionally, we explored the integration of embeddings from the pre-trained model BEATs, an iterative framework to learn Bidirectional Encoder representation from Audio Transformers. By concatenating these embeddings into the input of the Conformer blocks, results were further improved, achieving a PSDS2 value of 0.813 and considerably outperforming SED systems based on CRNNs.

RESUMEN En este articulo presenta el diseño, implementación y comparación de técnicas de control activo de ruido, estos controladores se basa en técnicas de control adaptativo mediante un filtro FIR y el algoritmo LMS, el principio de funcionamiento de esta ...

The Pulse Code Modulation is the vital part of Analog to Digital Converter (ADC). The PCM includes the process of sampling and quantization in order to digitize the analog input along the time scale and amplitude scale respectively. This project aims to design the Pulse Code Modulation Multiplexer (PCMM) in which analog inputs with different frequencies are multiplexed and pulse code modulated. The PCMM is developed using Very High Speed Integrated Circuit Hardware Description Language (VHDL). The proposed Pulse Code Modulated Multiplexer is validated by comparing the design parameters like power and area, implemented in the Xilinx Spartan Field Programmable Gate Array (FPGA).

Based on perceptual and computational attention modeling studies, we formulate measures of saliency for an audiovisual stream. Audio saliency is captured by signal modulations and related multifrequency band features, extracted through nonlinear operators and energy tracking. Visual saliency is measured by means of a spatiotemporal attention model driven by various feature cues (intensity, color, motion). Audio and video curves are integrated in a single attention curve, where events may be enhanced, suppressed or vanished. The presence of salient events is signified on this audiovisual curve by geometrical features such as local extrema, sharp transition points and level sets. An audiovisual saliency-based movie summarization algorithm is proposed and evaluated. The algorithm is shown to perform very well in terms of summary informativeness and enjoyability for movie clips of various genres.

This document provides different tools and techniques useful for debugging audio issues in Linux, covering problems from booting to shutdown. This guide is not exhaustive but aims to explain potential audio issues or bugs that can arise when bringing up audio drivers on Linux or fixing existing audio corruption issues during playback or capture.

This paper investigates the contextual recognition of neutral thirds in music by integrating real-world musical context into the study of categorical perception. Traditionally, categorical perception has been studied using isolated auditory stimuli in controlled laboratory settings. However, music is typically experienced within a circumstantial framework, significantly influencing its reception. Our study involved musicians from various specializations who listened to precomposed musical fragments, each concluding with a 350-cent interval preceded by different harmonic contexts. The fragments included a monophonic synthesizer and orchestral mockups, with contexts such as major chords, minor chords, a single pitch, neutral thirds, and natural fifths. The results indicate that musical context remarkably affects the recognition of pseudotonal chords. Participants' accuracy in judging interval size varied based on the preceding harmonic context. Two statistical analyses was conducted to determine if there were significant differences in the neutral third perception across the different harmonic contexts (p = 9.86*10^-29) and across the cultural background of the study participants (p = 0.8998). The findings underscore the need to consider real-world listening experiences in research on auditory processing and cognition.

Auscultation is one of the most used techniques for detecting cardiovascular diseases, which is one of the main causes of death in the world. Heart murmurs are the most common abnormal finding when a patient visits the physician for auscultation. These heart sounds can either be innocent, which are harmless, or abnormal, which may be a sign of a more serious heart condition. However, the accuracy rate of primary care physicians and expert cardiologists when auscultating is not good enough to avoid most of both type-I (healthy patients are sent for echocardiogram) and type-II (pathological patients are sent home without medication or treatment) errors made. In this paper, the authors present a novel convolutional neural network based tool for classifying between healthy people and pathological patients using a neuromorphic auditory sensor for FPGA that is able to decompose the audio into frequency bands in real time. For this purpose, different networks have been trained with the heart murmur information contained in heart sound recordings obtained from nine different heart sound databases sourced from multiple research groups. These samples are segmented and preprocessed using the neuromorphic auditory sensor to decompose their audio information into frequency bands and, after that, sonogram images with the same size are generated. These images have been used to train and test different convolutional neural network architectures. The best results have been obtained with a modified version of the AlexNet model, achieving 97% accuracy (specificity: 95.12%, sensitivity: 93.20%, PhysioNet/CinC Challenge 2016 score: 0.9416). This tool could aid cardiologists and primary care physicians in the auscultation process, improving the decision making task and reducing type-I and type-II errors.

Speech recognition has become an important task to improve the human-machine interface. Taking into account the limitations of current automatic speech recognition systems, like non-real time cloud-based solutions or power demand, recent interest for neural networks and bio-inspired systems has motivated the implementation of new techniques. Among them, a combination of spiking neural networks and neuromorphic auditory sensors offer an alternative to carry out the human-like speech processing task. In this approach, a spiking convolutional neural network model was implemented, in which the weights of connections were calculated by training a convolutional neural network with specific activation functions, using firing rate-based static images with the spiking information obtained from a neuromorphic cochlea. The system was trained and tested with a large dataset that contains "left" and "right" speech commands, achieving 89.90% accuracy. A novel spiking neural network model has been proposed to adapt the network that has been trained with static images to a non-static processing approach, making it possible to classify audio signals and time series in real time.

Sound events possess certain temporal and spectral structure in their time-frequency representations. The spectral content for the samples of the same sound event class may exhibit small shifts due to intra-class acoustic variability. Convolutional layers can be used to learn high-level, shift invariant features from time-frequency representations of acoustic samples, while recurrent layers can be used to learn the longer term temporal context from the extracted high-level features. In this paper, we propose combining these two in a convolutional recurrent neural network (CRNN) for rare sound event detection. The proposed method is evaluated over DCASE 2017 challenge dataset of individual sound event samples mixed with everyday acoustic scene samples. CRNN provides significant performance improvement over two other deep learning based methods mainly due to its capability of longer term temporal modeling.

We present an extensive evaluation of a wide variety of promising design patterns for automated deep-learning (AutoDL) methods, organized according to the problem categories of the 2019 AutoDL challenges, which set the task of optimizing both model accuracy and search efficiency under tight time and computing constraints. We propose structured empirical evaluations as the most promising avenue to obtain design principles for deep-learning systems due to the absence of strong theoretical support. From these evaluations, we distill relevant patterns which give rise to neural network design recommendations. In particular, we establish (a) that very wide fully connected layers learn meaningful features faster; we illustrate (b) how the lack of pretraining in audio processing can be compensated by architecture search; we show (c) that in text processing deep-learning-based methods only pull ahead of traditional methods for short text lengths with less than a thousand characters under tig...

Negli ultimi anni si è assistito a un incredibile aumento del numero di smartcard prodotte sul mercato; è sufficiente pensare alla quantità di SIM card presenti nel mondo per farsi un’idea del volume complessivo di vendite che nel solo anno 2008 ha raggiunto un valore superiore ai cinque miliardi di unità. Nonostante la definizione di un numero significativo di standard per regolare l’utilizzo e il funzionamento di questi dispositivi, lo sviluppo di soluzioni basate su smartcard è ancora notevolmente complesso. Una delle cause principali di questa complessità è la difficoltà nel testare le proprie applicazioni in modo automatico che è dovuta principalmente all’assenza di strumenti adatti a tale scopo. La maggior parte degli strumenti disponibili sono proprietari e utilizzabili solamente attraverso un’interfaccia grafica che rende di fatto impossibile l’automatizzazione dei test. I pochi strumenti che lo consentono, tuttavia, sono fortemente non interoperabili per ragioni puramente c...

RESUMEN En este articulo presenta el diseño, implementación y comparación de técnicas de control activo de ruido, estos controladores se basa en técnicas de control adaptativo mediante un filtro FIR y el algoritmo LMS, el principio de funcionamiento de esta ...

Auscultation is one of the most used techniques for detecting cardiovascular diseases, which is one of the main causes of death in the world. Heart murmurs are the most common abnormal finding when a patient visits the physician for auscultation. These heart sounds can either be innocent, which are harmless, or abnormal, which may be a sign of a more serious heart condition. However, the accuracy rate of primary care physicians and expert cardiologists when auscultating is not good enough to avoid most of both type-I (healthy patients are sent for echocardiogram) and type-II (pathological patients are sent home without medication or treatment) errors made. In this paper, the authors present a novel convolutional neural network based tool for classifying between healthy people and pathological patients using a neuromorphic auditory sensor for FPGA that is able to decompose the audio into frequency bands in real time. For this purpose, different networks have been trained with the heart murmur information contained in heart sound recordings obtained from nine different heart sound databases sourced from multiple research groups. These samples are segmented and preprocessed using the neuromorphic auditory sensor to decompose their audio information into frequency bands and, after that, sonogram images with the same size are generated. These images have been used to train and test different convolutional neural network architectures. The best results have been obtained with a modified version of the AlexNet model, achieving 97% accuracy (specificity: 95.12%, sensitivity: 93.20%, PhysioNet/CinC Challenge 2016 score: 0.9416). This tool could aid cardiologists and primary care physicians in the auscultation process, improving the decision making task and reducing type-I and type-II errors.

A graphic equalizer is an adjustable filter in which the command gain of each frequency band is practically independent of the gains of other bands. Designing a graphic equalizer with a high precision requires evaluating a target response that interpolates the magnitude response at several frequency points between the command gains. Good accuracy has been previously achieved by using polynomial interpolation methods such as cubic Hermite or spline interpolation. However, these methods require large computational resources, which is a limitation in real-time applications. This paper proposes an efficient way of computing the target response without sacrificing the approximation accuracy. This new approach called Linear Interpolation with Constant Segments (LICS) reduces the computing time of the target response by 55% and has an intrinsic parallel structure. Performance of the LICS method is assessed on an ARM Cortex-A7 core, which is commonly used in embedded systems.

We address the problem of blind audio source separation in the under-determined and convolutive case. The contribution of each source to the mixture channels in the time-frequency domain is modeled by a zero-mean Gaussian random vector with a full rank covariance matrix composed of two terms: a variance which represents the spectral properties of the source and which is modeled by a nonnegative matrix factorization (NMF) model and another full rank covariance matrix which encodes the spatial properties of the source contribution in the mixture. We address the estimation of these parameters by maximizing the likelihood of the mixture using an expectation-maximization (EM) algorithm. Theoretical propositions are corroborated by experimental studies on stereo reverberant music mixtures.

In this paper we propose a novel time-space ensemble-based approach for the task of automatic music genre classification. Ensemble strategies employ several classifiers to different views of the problemspace, and combination rules in order to produce the final classification decision. In our approach we employ audio signal segmentation in time intervals and also problem space decomposition. Initially the music signal is split in time segments; features are extracted from these music signal segments and the one against all (OAA) and round robin (RR) strategies, which implement a space decomposition by using several binary classifiers, are applied. Finally, the outputs of the set of classifiers are combined to produce the final result. We test our proposition in a music database of 1.200 music samples from four different music genres. Experimental results show that time segment decomposition is more important than the space decomposition produced by the OAA and RR strategies, although they produce better results relative to the use of single classifiers and feature vectors.

In questa tesi viene proposta e sperimentata una metodologia per il supporto di applicazioni grid-aware espresse come moduli paralleli ASSIST puri e incapsulati in componenti CCM (CORBA Component Model). La metodologia oggetto della tesi prevede il monitoraggio del tempo di servizio di applicazioni, basate su stream di dati, e la riconfigurazione da parte di un Application Manager, che può autonomamente decidere quali moduli/componenti paralleli utilizzare per bilanciare il carico, rispettando quindi il contratto di performance o evitando uno spreco inutile di risorse. A conclusione della sperimentazione, viene fatto un confronto tra i due approcci proposti (ASSIST puro ed integrato CCM), valutando l'impatto dei componenti CCM in una applicazione in esecuzione su una griglia, e valutando la bontà dell'Application Manager nei due casi

Recently, there is an increasing tendency to embed functionalities for recognizing emotions from user generated media content in automated systems such as call-centre operations, recommendations and assistive technologies, providing richer and more informative user and content profiles. However, to date, adding these functionalities was a tedious, costly, and time consuming effort, requiring identification and integration of diverse tools with diverse interfaces as required by the use case at hand. The MixedEmotions Toolbox leverages the need for such functionalities by providing tools for text, audio, video, and linked data processing within an easily integrable plug-and-play platform. These functionalities include: (i) for text processing: emotion and sentiment recognition, (ii) for audio processing: emotion, age, and gender recognition, (iii) for video processing: face detection and tracking, emotion recognition, facial landmark localization, head pose estimation, face alignment, and body pose estimation, and (iv) for linked data: knowledge graph integration. Moreover, the MixedEmotions Toolbox is open-source and free. In this article, we present this toolbox in the context of the existing landscape, and provide a range of detailed benchmarks on standard test-beds showing its state-of-the-art performance. Furthermore, three real-world use-cases show its effectiveness, namely emotion-driven smart TV, call center monitoring, and brand reputation analysis.

This contribution proposes an efficient method for the detection ofrelevant changes in continuous stream of sound. The detectedchange-points can then serve for the segmentation of long audiorecordings into shorter and more or less homogenous sections. First, wediscuss the task of a single change-point detection using the Bayesdecision theory. We show that it leads to a quite simple andcomputationally efficient solution based on the Bayesian InformationCriterion. Next, we extend this approach to formulate the algorithm forthe detection of multiple change-points. Finally, the proposedalgorithm is applied for the segmentation of broadcast newsaudio-streams into parts belonging to different speakers or differentacoustic conditions. Such segmentation is necessary as the first stepin the automatic speech-to-text transcription of TV or radio news.

Close-microphone techniques are extensively employed in many live music recordings, allowing for interference rejection and reducing the amount of reverberation in the resulting instrument tracks. However, despite the use of directional microphones, the recorded tracks are not completely free from source interference, a problem which is commonly known as microphone leakage. While source separation methods are potentially a solution to this problem, few approaches take into account the huge amount of prior information available in this scenario. In fact, besides the special properties of close-microphone tracks, the knowledge on the number and type of instruments making up the mixture can also be successfully exploited for improved separation performance. In this paper, a nonnegative matrix factorization (NMF) method making use of all the above information is proposed. To this end, a set of instrument models are learnt from a training database and incorporated into a multichannel extension of the NMF algorithm. Several options to initialize the algorithm are suggested, exploring their performance in multiple music tracks and comparing the results to other state-of-the-art approaches.

We address the problem of identifying the constituent sources in a single-sensor mixture signal consisting of contributions from multiple simultaneously active sources. We propose a generic framework for mixture signal analysis based on a latent variable approach. The basic idea of the approach is to detect known sources represented as stochastic models, in a single-channel mixture signal without performing signal separation. A given mixture signal is modeled as a convex combination of known source models and the weights of the models are estimated using the mixture signal. We show experimentally that these weights indicate the presence/absence of the respective sources. The performance of the proposed approach is illustrated through mixture speech data in a reverberant enclosure. For the task of identifying the constituent speakers using data from a single microphone, the proposed approach is able to identify the dominant source with up to 8 simultaneously active background sources in a room with RT60 250 ms, using models obtained from clean speech data for a Source to Interference Ratio (SIR) greater than 2 dB.

Finally, I would like to thank all my friends and family who tried to understand and respect my choice, and who never tried to prevent me from doing this "crazy" jump into the void. I am coming back better and stronger, and this is all thanks to you, because you let that happen. Despite the thousands of miles between us, the love you sent me, day after day, was received straight into my heart, and kept me standing up even when I felt down. Thank you so very much to my friends

The develop of Antescofo software has allowed contemporary musicians to create interactive music pieces in a more precise way in terms of the synchronization between human and machine. INRIA's MUTANT team has been developing a version of Antescofo that changes the DSP computation paradigm and translates the responsibility to the software and no longer to the host environment. Thus, the composer gets more freedom to create his own effects. Plus, it allows the composer to change the sound link network in execution time. Also, the computational power required for the DSP is optimized. Lastly, this new version creates the possibility of generating a self-contained score for UDOO platform, creating this way the first steps toward the preservation of interactive musical pieces through time. In this article the context of the new version of Antescofo is presented, as well as the benefits of it. The methodology of compilation of the Faust DSP tool in Antescofo is described and a comparison of the profiling tests between the old and new versions is detailed. The results shows an improvement of about 46% in terms of the computational power needed to process the signals and they represent clear indicators of optimization that can be extended to the compilation of the Antescofo software for the UDOO platform.

A recent trend is to use Music Information Retrieval algorithms for creativity. When considering the audio signal as observation, a well-known method of data-driven synthesis is the "concatenative synthesis" also named musaicing (audio mosaicing) [7]. We propose here to extend the musaicing framework by including recent advances in audio source separation. We propose a method which allows reproducing a target music track by superimposing in time and frequency constitutive elements obtained using source separation methods. 1

A swarm of bees buzzing "Let it be" by the Beatles or the wind gently howling the romantic "Gute Nacht" by Schubert-these are examples of audio mosaics as we want to create them. Given a target and a source recording, the goal of audio mosaicing is to generate a mosaic recording that conveys musical aspects (like melody and rhythm) of the target, using sound components taken from the source. In this work, we propose a novel approach for automatically generating audio mosaics with the objective to preserve the source's timbre in the mosaic. Inspired by algorithms for non-negative matrix factorization (NMF), our idea is to use update rules to learn an activation matrix that, when multiplied with the spectrogram of the source recording, resembles the spectrogram of the target recording. However, when applying the original NMF procedure, the resulting mosaic does not adequately reflect the source's timbre. As our main technical contribution, we propose an extended set of update rules for the iterative learning procedure that supports the development of sparse diagonal structures in the activation matrix. We show how these structures better retain the source's timbral characteristics in the resulting mosaic.

Cough efficacy is considered a reliable predictor of the aspiration risk in head and neck cancer patients with radiation-associated dysphagia. Currently, coughing is assessed perceptually or aerodynamically. The goal of our research is to develop methods of acoustic cough analysis. In this study, we examined in a healthy population the acoustical differences between three protective maneuvers: voluntary cough, voluntary throat clearing, and induced reflexive cough. Forty healthy participants were included in this study. Voluntary cough, voluntary throat clearing, and reflexive cough samples were recorded and analyzed acoustically. Temporal acoustic features were the following: the slope and curvature of the amplitude contour, as well as the average, slope, and curvature of the sample entropy and kurtosis contours of the recorded signal. Spectral features were the relative energy in the frequency bands (0–400 Hz, 400–800 Hz, 800–1600 Hz, 1600 Hz-3200 Hz, > 3200 Hz) as well as the ...

Music genre classification has its own popularity index in the present times. Machine learning can play an important role in the music streaming task. This research article proposes a machine learning based model for the classification of music genre. The evaluation of the proposed model is carried out while considering different music genres as in blues, metal, pop, country, classical, disco, jazz and hip-hop. Different audio features utilized in this study include MFCC (Mel Frequency Spectral Coefficients), Delta, Delta-Delta and temporal aspects for processing the data. The implementation of the proposed model has been done in the Python language. The results of the proposed model reveal an accuracy SVM accuracy of 95%. The proposed algorithm has been compared with existing algorithms and the proposed algorithm performs better than the existing ones in terms of accuracy.

This paper introduces a novel declipping algorithm based on constrained least-squares minimization. Digital speech signals are often sampled at 16 kHz and classic declipping algorithms fail to accurately reconstruct the signal at this sampling rate due to the scarcity of reliable samples after clipping. The Constrained Blind Amplitude Reconstruction algorithm interpolates missing data points such that the resulting function is smooth while ensuring the inferred data fall in a legitimate range. The inclusion of explicit constraints helps to guide an accurate interpolation. Evaluation of declipping performance is based on automatic speech recognition word error rate and Constrained Blind Amplitude Reconstruction is shown to outperform the current state-of-the-art declipping technology under a variety of conditions. Declipping performance in additive noise is also considered.

Computer-assisted audio recordings provide a new approach for detecting and correcting interviewer coding error. For questions with categorical and “other specify” responses, it is possible for the interviewer to misinterpret, abbreviate, or improperly key the respondent's answer. ...

This paper investigates the use of a physical model template database as the parameter basis for a MPEG-4 Structured Audio (MP4-SA) codec. During analysis, the codec attempts to match the closest corresponding instrument in the database. In this paper, we emphasize the mechanism enabling this match. We give an overview of the final front end, including the pitch detection stage, and remaining problems are discussed. A draft implementation, written in the Python language is described.

While humans can act as effective sensors, human input is subject to a high degree of error and highly dependent on the context. Furthermore, extracting the signal from the noise for social sensing is a difficult challenge. One approach to improving the accuracy of social sensing is to use physical sensors as a control knob for social sensing algorithms. In this paper, we present an architecture for using audio sensors as a way to control an algorithm used for social sensing of interesting events. We present various use cases where the architecture is applicable, and go into the details of one specific use case, namely using crowd behavior in a golf-course to identify and control social media feeds related to the course. CCS CONCEPTS •Human-centered computing →Collaborative and social computing; •Computer systems organization →Distributed architectures; •Computing methodologies →Machine learning approaches;

Multivariate Analysis (MVA) comprises a family of well-known methods for feature extraction which exploit correlations among input variables representing the data. One important property that is enjoyed by most such methods is uncorrelation among the extracted features. Recently, regularized versions of MVA methods have appeared in the literature, mainly with the goal to gain interpretability of the solution. In these cases, the solutions can no longer be obtained in a closed manner, and more complex optimization methods that rely on the iteration of two steps are frequently used. This paper recurs to an alternative approach to solve efficiently this iterative problem. The main novelty of this approach lies in preserving several properties of the original methods, most notably the uncorrelation of the extracted features. Under this framework, we propose a novel method that takes advantage of the 2,1 norm to perform variable selection during the feature extraction process. Experimental results over different problems corroborate the advantages of the proposed formulation in comparison to state of the art formulations.

Wireless sensor network (WSN) has proliferated rapidly as a cost-effective solution for data aggregation and measurements under challenging environments. Sensors in WSNs are cheap, powerful, and consume limited energy. The energy consumption is considered to be the dominant concern because it has a direct and significant influence on the application's lifetime. Recently, the availability of small and inexpensive components such as microphones has promoted the development of wireless acoustic sensor networks (WASNs). Examples of WASN applications are hearing aids, acoustic monitoring, and ambient intelligence. Monitoring animals, especially those that are becoming endangered, can assist with biology researchers' preservation efforts. In this work, we first focus on exploring the existing methods used to monitor the animal by recognizing their sounds. Then we propose a new energy-efficient approach for identifying animal sounds based on the frequency features extracted from acoustic sensed data. This approach represents a suitable solution that can be implemented and used in various applications. However, the proposed system considers the balance between application efficiency and the sensor's energy capabilities. The energy savings will be achieved through processing the recognition tasks in each sensor, and the recognition results will be sent to the base station.

While previous generations of the MPEG multimedia standard have focused primarily on coding and transmission of content digitally sampled from the real world, MPEG-4 contains extensive support for structured, synthetic and synthetic/natural hybrid coding methods. An overview is presented of the "Structured Audio" and "AudioBIFS" components of MPEG-4, which enable the description of synthetic soundtracks, musical scores, and effects algorithms and the compositing, manipulation, and synchronization of real and synthetic audio sources. A discussion of the separation of functionality between the systems layer and the audio toolset of MPEG-4 is presented, and prospects for efficient DSP-based implementations are discussed.

We present a new method for the discrimination of explosive cough events, which is based on a combination of spectral content descriptors and pitch-related features. After the removal of near-silent segments, a vector of event boundaries is obtained and a proposed set of 9 features is extracted for each event. Two data sets, recorded using electronic stethoscopes and comprising a total of 46 healthy subjects and 13 patients, were employed to evaluate the method. The proposed feature set is compared to three other sets of descriptors: a baseline, a combination of both sets, and an automatic selection of the best 10 features from both sets. The combined feature set yields good results on the cross-validated database, attaining a sensitivity of 92.32.3% and a specificity of 84.73.3%. Besides, this feature set seems to generalize well when it is trained on a small data set of patients, with a variety of respiratory and cardiovascular diseases, and tested on a bigger data set of mostly healthy subjects: a sensitivity of 93.4% and a specificity of 83.4% are achieved in those conditions. These results demonstrate that complementing the proposed feature set with a baseline set is a promising approach.

The problem of modeling a signal segment as a sum of exponentially damped sinusoidal components arises in many different application areas, including speech and audio processing. Often, model parameters are estimated using subspace based techniques which arrange the input signal in a structured matrix and exploit the so-called shift-invariance property related to certain vector spaces of the input matrix. A problem with this class of estimation algorithms, when used for speech and audio processing, is that the perceptual importance of the sinusoidal components is not taken into account. In this work we propose a solution to this problem. In particular, we show how to combine well-known subspace based estimation techniques with a recently developed perceptual distortion measure, in order to obtain an algorithm for extracting perceptually relevant model components. In analysis-synthesis experiments with wideband audio signals, objective and subjective evaluations show that the proposed algorithm improves perceived signal quality considerable over traditional subspace based analysis methods.

In the past few years, several case studies have illustrated that the use of occupancy information in buildings leads to energy-efficient and low-cost HVAC operation. The widely presented techniques for occupancy estimation include temperature, humidity, CO2 concentration, image camera, motion sensor and passive infrared (PIR) sensor. So far little studies have been reported in literature to utilize audio and speech processing as indoor occupancy prediction technique. With rapid advances of audio and speech processing technologies, nowadays it is more feasible and attractive to integrate audio-based signal processing component into smart buildings. In this work, we propose to utilize audio processing techniques (i.e., speaker recognition and background audio energy estimation) to estimate room occupancy (i.e., the number of people inside a room). Theoretical analysis and simulation results demonstrate the accuracy and effectiveness of this proposed occupancy estimation technique. Based on the occupancy estimation, smart buildings will adjust the thermostat setups and HVAC operations, thus, achieving greater quality of service and drastic cost savings.

In this paper, we describe our multi-resolution mean teacher systems for DCASE 2021 Task 4: Sound event detection and separation in domestic environments. Aiming to take advantage of the different lengths and spectral characteristics of each target category, we follow the multi-resolution feature extraction approach that we introduced for last year's edition. It is found that each one of the proposed Polyphonic Sound Detection Score (PSDS) scenarios benefits from either a higher temporal resolution or a higher frequency resolution. Additionally, the combination of several time-frequency resolutions through model fusion is able to improve the PSDS results in both scenarios. Furthermore, a class-wise analysis of the PSDS metrics is provided, indicating that the detection of each event category is optimized with different resolution points or model combinations.

Cough detection and classification present necessary tools for the evaluation of pathology severity in chronic illnesses. In literature, several approaches have been proposed for this aim. The latter presented a relative success since none of them allow a standardized exploitation and interpretation of sensors results in order to provide an efficient cough detection and classification. This paper presents a new system that disposes of a prototype (hardware) for data acquisition, and software for exploiting the acquired data for cough detection, visualization and classification. Our prototype includes sensors such as ECG, thermistor, chest belt, accelerometer, oximeter, contact and audio microphones. The relevance of each sensor is evaluated within three features: mutual information obtained with the features, ability to distinguish cough from other event categories, and ability to detect cough events. The sensors values are interpreted and visualized with a graphical view where the ...

Cough efficacy is considered a reliable predictor of the aspiration risk in head and neck cancer patients with radiation-associated dysphagia. Currently, coughing is assessed perceptually or aerodynamically. The goal of our research is to develop methods of acoustic cough analysis. In this study, we examined in a healthy population the acoustical differences between three protective maneuvers: voluntary cough, voluntary throat clearing, and induced reflexive cough. Forty healthy participants were included in this study. Voluntary cough, voluntary throat clearing, and reflexive cough samples were recorded and analyzed acoustically. Temporal acoustic features were the following: the slope and curvature of the amplitude contour, as well as the average, slope, and curvature of the sample entropy and kurtosis contours of the recorded signal. Spectral features were the relative energy in the frequency bands (0–400 Hz, 400–800 Hz, 800–1600 Hz, 1600 Hz-3200 Hz, > 3200 Hz) as well as the ...

Cough efficacy is considered a reliable predictor of the aspiration risk in head and neck cancer patients with radiation-associated dysphagia. Currently, coughing is assessed perceptually or aerodynamically. The goal of our research is to develop methods of acoustic cough analysis. In this study, we examined in a healthy population the acoustical differences between three protective maneuvers: voluntary cough, voluntary throat clearing, and induced reflexive cough. Forty healthy participants were included in this study. Voluntary cough, voluntary throat clearing, and reflexive cough samples were recorded and analyzed acoustically. Temporal acoustic features were the following: the slope and curvature of the amplitude contour, as well as the average, slope, and curvature of the sample entropy and kurtosis contours of the recorded signal. Spectral features were the relative energy in the frequency bands (0–400 Hz, 400–800 Hz, 800–1600 Hz, 1600 Hz-3200 Hz, > 3200 Hz) as well as the ...

A fast iterative Kernel Principal Component Analysis (KPCA) is proposed to extract features from hyperspectral images. The proposed method is a kernel version of the Candid Covariance-Free Incremental Principal Component Analysis, which solves the eigenvectors through iteration. Without performing eigen decomposition on Gram matrix, our method can reduce the space complexity and time complexity greatly. Experimental results were validated in comparison with the standard KPCA and linear version methods.

This abstract describes the tempo extraction algorithm used for the University of Victoria submission to the MIREX (Music Information Retrieval Exchange) 2005. The algorithm is mostly based on self-similarity rather than onset detection. However, an onset detection component is used to calculate the phase of the dominant periodicities. Multiple frequency bands are calculated using a Discrete Wavelet Transform. Subsequently the envelope of each band is extracted and autocorrelation is used to find the dominant ...