No description available.

The main focus of the research presented within this thesis is, as the title suggests, binaural spatialization. Binaural technology and, especially, the binaural recording technique are not particu-larly recent. Nevertheless, the interest in this technology has lately become substantial due to the increase in the calculation power of personal computers, which started to allow the complete and accurate real-time simulation of three-dimensional sound-fields over headphones. The goals of this body of research have been determined in order to provide elements of novelty and of contribution to the state of the art in the field of binaural spatialization. A brief summary of these is found in the following list: • The development and implementation of a binaural spatialization technique with Distance Simulation, based on the individual simulation of the distance cues and Binaural Reverb, in turn based on the weighted mix between the signals convolved with the different HRIR and BRIR sets; • The development and implementation of a characterization process for modifying a BRIR set in order to simulate different environments with different characteristics in terms of frequency response and reverb time; • The creation of a real-time and offline binaural spatialization application, imple-menting the techniques cited in the previous points, and including a set of multichannel(and Ambisonics)-to-binaural conversion tools. • The performance of a perceptual evaluation stage to verify the effectiveness, realism, and quality of the techniques developed, and • The application and use of the developed tools within both scientific and artistic “case studies”. In the following chapters, sections, and subsections, the research performed between January 2006 and March 2010 will be described, outlining the different stages before, during, and after the development of the software platform, analysing the results of the perceptual evaluations and drawing conclusions that could, in the future, be considered the starting point for new and innovative research projects.

In hands-free telephony and in teleconference systems, the main aim is to provide agood free voice quality when two or more people communicate from different places.The problem often arises during the conversation is the creation of acoustic echo. Thisproblem will cause the bad quality of voice signal and thus talkers could not hearclearly the content of the conversation, even thought lost the important information.This acoustic echo is actually the noise which is created by the reflection of soundwaves by the wall of the room and the other things exist in the room. The mainobjective for engineers is the cancellation of this acoustic echo and provides an echofree environment for speakers during conversation. For this purpose, scientists designdifferent adaptive filter algorithms. Our thesis is also to study and simulate theacoustics echo cancellation by using different adaptive algorithms.

<p>In hands-free telephony and in teleconference systems, the main aim is to provide agood free voice quality when two or more people communicate from different places.The problem often arises during the conversation is the creation of acoustic echo. Thisproblem will cause the bad quality of voice signal and thus talkers could not hearclearly the content of the conversation, even thought lost the important information.This acoustic echo is actually the noise which is created by the reflection of soundwaves by the wall of the room and the other things exist in the room. The mainobjective for engineers is the cancellation of this acoustic echo and provides an echofree environment for speakers during conversation. For this purpose, scientists designdifferent adaptive filter algorithms. Our thesis is also to study and simulate theacoustics echo cancellation by using different adaptive algorithms.</p>

The hospital sound environment is complex. Emergency Departments (EDs), in particular, have proven to be hectic work environments populated with diverse sound sources. Medical equipment, alarms, and communication events generate noise that can interfere with staff concentration and communication. In this study, sound measurements and analyses were conducted in six hospitals total: three civilian hospitals in Atlanta, Georgia and Dublin, Ohio, as well as three Washington, DC-area hospitals in the Military Health System (MHS). The equivalent, minimum, and maximum sound pressure levels were recorded over twenty-four hours in several locations in each ED, with shorter 15-30 minute measurements performed in other areas. Acoustic descriptors, such as spectral content, level distributions, and speech intelligibility were examined. The perception of these acoustic qualities by hospital staff was also evaluated through subjective surveys. It was found that noise levels in both work areas and patient rooms were excessive. Additionally, speech intelligibility measurements and survey results show that background noise presents a significant obstacle in effective communication between staff members and patients. Compared to previous studies, this study looks at a wider range of acoustic metrics and the corresponding perceptions of staff in order to form a more precise and accurate depiction of the ED sound environment.

Tato diplomová práce se zabývá problematikou vzdáleného rozpoznávání mluvčích. V případě dat zachycených odlehlým mikrofonem se přesnost standardního rozpoznávání značně snižuje, proto jsem navrhl dva přístupy pro zlepšení výsledků. Prvním z nich je použití mikrofonního pole (záměrně rozestavené sady mikrofonů), které je schopné nasměrovat virtuální "paprsek" na pozici řečníka. Dále jsem prováděl adaptaci komponent systému (PLDA skórování a extraktoru i-vektorů). S využitím simulace pokojových podmínek jsem syntetizoval trénovací a testovací data ze standardní datové sady NIST 2010. Ukázal jsem, že obě techniky a jejich kombinace vedou k výraznému zlepšení výsledků. Dále jsem se zabýval společným určením identity a pozice mluvčího. Zatímco výsledky ve venkovním simulovaném prostředí (bez ozvěn) jsou slibné, výsledky z interiéru (s ozvěnami) jsou smíšené a vyžadují další prozkoumání. Na závěr jsem mohl systémem vyhodnotit omezené množství reálných dat získaných přehráním a záznamem nahrávek ve skutečné místnosti. Zatímco výsledky pro mužské nahrávky odpovídají simulaci, výsledky pro ženské nahrávky nejsou přesvědčivé a vyžadují další analýzu.

In Virtual Reality (VR) systems, the incorporation of acoustics allows for the generation of audio-visual stimuli, facilitating applications in engineering, architecture, and design. The goal of virtual acoustics is to create a realistic sound field in continuous space. Realistic virtual acoustic environments can be produced with wave-based acoustic simulations. However, rendering a sound field with a dense grid of room impulse responses (RIRs) in real-time is slow and memory-intensive. Conventionally, a more sparsely spaced grid of RIRs is used and as a workaround linear interpolation between the nearest RIRs is performed, allowing users to listen at an arbitrary location. However, the linear interpolation method reduces the quality of the sound field as it does not produce natural-sounding RIRs. The aim of this thesis is therefore to answer the question of whether we are able to achieve a better interpolation technique than linear interpolation using a machine learning approach. In this thesis, we present a novel neural network-based method for interpolating between Room Impulse Responses (RIRs). The networks were trained using RIRs from a wave-based simulation of a single 3D room and developed through a series of experiments. The experimental process was performed in three distinct stages. Firstly, we explored various representations of the RIRs: unprocessed RIRs, Short-time Fourier transform (STFT) of RIRs, and encoded STFT of the RIRs using an autoencoder. Secondly, we examined several different neural network architectures: Multi-layer perception, residual neural network, autoencoder, and U-Net. Additionally, we experimented with training the networks in a Generative Adversary Network (GAN) setting. Thirdly, we experimented with different sizes of the best-performing architecture. Results show that using an STFT representation of the RIRs combined with a residual neural network architecture yielded the most optimal results. Furthermore, we were able to outperform the established linear interpolation baseline. / Inom Virtuell Verklighet (VR) möjliggör användningen av akustik skapandet av audiovisuell stimuli, vilket underlättar tillämpningar inom ingenjörsvetenskap, arkitektur och design. Målet med virtuell akustik är att skapa ett verklighetstroget och kontinuerligt ljudfält. Verklighetstrogna virtuella akustiska miljöer kan skapas med hjälp av vågbaserade akustiska simuleringar. Men att återge ett ljudfält med ett tätt rutnät av Room Impulse Responses (RIRs) i realtid är långsamt och minneskrävande. Konventionellt används ett rutnät med glesare avstånd av RIR, och som en lösning utförs linjär interpolation mellan de närmaste RIR:erna, vilket tillåter användare att lyssna på en godtycklig plats. Den linjära interpolationen minskar dock kvaliteten på ljudfältet eftersom den inte producerar naturligt ljudande RIR:er. Syftet med detta examensarbete är därför att besvara frågan om vi kan finna en bättre interpolationsteknik än linjär interpolation med hjälp av en maskininlärningsmetod. I detta examensarbete presenterar vi en ny metod för interpolering mellan Room Impulse Responses (RIR:er) baserad på neurala nätverk. De neurala nätverken tränades med hjälp av RIR:er från en vågbaserad simulering av ett enskilt 3D-rum och utvecklades genom en serie experiment. Experimenten utfördes i tre steg. Först undersöktes olika representationer av RIR:er: obearbetade RIR:er, korttids fouriertransform (STFT) av RIR:er och kodade STFT av RIR:er med hjälp av en autoencoder. Det andra steget innefattade undersökningen av flera olika neurala nätverksarkitekturer: Multi-layer perception, residual neural network, autoencoder och U-Net. Dessutom experimenterade vi med att träna nätverken i en GAN-miljö (Generative Adversary Network). I det tredje steget experimenterade vi med olika storlekar på den mest effektiva arkitekturen. Resultaten visar att användning av en STFT-representation av RIR:er kombinerat med en residual neural nätverksarkitektur resulterade i de mest optimala resultaten. Dessutom kunde vi överträffa den etablerade linjära interpolationsbaslinjen.

No description available.

The main focus of this thesis is to analyse signals (signal-dependent analysis) and room responses (system-dependent analysis) from a statistical point of view, attempt to determine the underlying statistical relationships between the reverberant signals and the room responses and propose relevant statistical models. Based on such a statistical framework, this thesis aims to propose novel methodologies for the extraction of room acoustical information and parameters from reverberant signals. Schroeder's theory is experimentally evaluated for various Room Transfer Functions (RTFs) measured in many source/receiver positions in various enclosures and several related aspects are discussed. Using a statistical approach, the effects of reverberant energy on the histograms and statistical measures are discussed and models describing the relationship of statistical measures between the reverberant signal and the RTFs are extracted. Then, the statistical properties of Binaural Room Transfer Functions (BRTFs) and binaural cues are examined. The well-known property of the spectral standard deviation of the magnitude of RTFs, that is its convergence to 5.6 dB for diffuse fields, is examined for the case of BRTFs, using a similar approach and a generic model for the relationship of the spectral standard deviation of RTFs and BRTFs. This thesis is also concerned with the distance estimation problem from a perceptual and computational point of view. Two novel methods for the estimation of the source/receiver distance using speech signals are proposed. The first method is able to detect the distance between the speaker and the microphone in a room environment using single-channel signals. The distance-dependent variation of several temporal and spectral statistical features of single-channel signals is studied and a novel sound source distance detector, based on these features is developed. The second method estimates distance from binaural speech signals (two-channel signals). This method does not require a priori knowledge of the room impulse response, the reverberation time or any other acoustical parameter and relies on a set of novel features extracted from the reverberant binaural signals. For this method, a novel distance estimation feature is introduced exploiting the standard deviation of the difference of the magnitude spectra of the left and right binaural signals (termed here as Binaural Spectral Magnitude Difference Standard Deviation (BSMD STD)). Moreover, an extended and novel set of additional features based on the statistical properties of binaural cues (ILDs, ITDs, ICs) is extracted from an auditory front-end which models the peripheral processing of the human auditory system. Both methods rely on novel distance-dependent features, related to statistical parameters of speech signals. Finally, a novel method for the estimation of the direct-to-reverberant-ratio (DRR) from dual-channel microphone recordings without having knowledge of the source signal is presented. / Η παρούσα διατριβή ασχολείται με τη μελέτη και ανάλυση των στατιστικών χαρακτηριστικών ηχητικών σημάτων και των ακουστικών αποκρίσεων χώρου, έχοντας ως πρωταρχικό σκοπό να προτείνει σχέσεις που περιγράφουν τη συσχέτιση των στατιστικών χαρακτηριστικών των σημάτων με αντήχηση με τις ακουστικές αποκρίσεις χώρων. Βάσει ενός τέτοιου θεωρητικού πλαισίου, η διατριβή αυτή αποσκοπεί στο να προτείνει νέες μεθοδολογίες για την εξαγωγή πληροφορίας που σχετίζεται με τα ακουστικά χαρακτηριστικά των χώρων, κάνοντας χρήση ηχογραφημένων ηχητικών σημάτων (π.χ. σήματα ομιλίας) στους εκάστοτε κλειστούς χώρους. Το θεωρητικό υπόβαθρο αυτής της διατριβής βασίζεται σε υπάρχοντα θεωρητικά μοντέλα για το ηχητικό πεδίο μέσα σε ένα κλειστό χώρο, όπως, για παράδειγμα, το στατιστικό μοντέλο του Schroeder. Το μοντέλο του Schroeder επιβεβαιώνεται πειραματικά για ακουστικές αποκρίσεις που έχουν μετρηθεί σε διάφορες θέσεις, μέσα σε κλειστούς χώρους, οι οποίοι διαφέρουν στα ακουστικά χαρακτηριστικά τους. Βάσει στατιστικής ανάλυσης, εξάγονται στατιστικά μοντέλα, τα οποία περιγράφουν την επίδραση της αντήχησης στα ηχητικά σήματα, όταν αυτά αναπαραχθούν μέσα σε ένα κλειστό χώρο. Στη συνέχεια, λαμβάνοντας υπόψη αντιληπτικά μοντέλα ακοής, τα οποία προϋποθέτουν την ύπαρξη δυο ηχητικών σημάτων (δυο αυτιά, αμφιωτική ακοή) σε αυτή τη διατριβή, μελετώνται κάποιες παράμετροι οι οποίες εξάγονται από αμφιωτικές ακουστικές αποκρίσεις χώρου. Η ιδιότητα της φασματικής τυπικής απόκλισης συναρτήσεων μεταφοράς χώρων να συγκλίνει στην τιμή των 5.6~dB για διάχυτα ηχητικά πεδία, επεκτείνεται στην περίπτωση των αμφιωτικών αποκρίσεων χώρου και προτείνεται ένα γενικευμένο μοντέλο που συσχετίζει τη φασματική τυπική απόκλιση μονοφωνικών και αμφιωτικών συναρτήσεων μεταφοράς χώρου. Η διατριβή αυτή, επίσης, ασχολείται με το πρόβλημα της εκτίμησης της απόστασης μεταξύ πηγής και δέκτη. Προτείνονται δυο νέες μέθοδοι για την εκτίμηση της απόστασης μεταξύ πηγής και δέκτη, κάνοντας χρήση ηχητικών σημάτων ομιλίας. Η προτεινόμενη μέθοδος βασίζεται σε μια σειρά από στατιστικές παραμέτρους των οποίων οι τιμές μεταβάλλονται είτε στο πεδίο του χρόνου είτε στο πεδίο της συχνότητας. Η δεύτερη προτεινόμενη μέθοδος αφορά, επίσης, στην εκτίμηση της απόστασης πηγής/δέκτη, αλλά από αμφιωτικά σήματα. Η μέθοδος αυτή δεν προαπαιτεί γνώση της ακουστικής απόκρισης του χώρου, του χρόνου αντήχησης ή άλλης ακουστικής παραμέτρου και βασίζεται σε μια σειρά από νέες παραμέτρους, οι οποίες μπορούν να υπολογισθούν από τα αμφιωτικά σήματα με αντήχηση. Οι παράμετροι συνδυάζονται με δυο διαφορετικές τεχνικές αναγνώρισης προτύπων των οποίων τα μειονεκτήματα και πλεονεκτήματα συζητώνται. Στα πλαίσια αυτής της μεθόδου, προτείνεται μια νέα παράμετρος, η οποία βασίζεται στη διαφορά της φασματικής τυπικής απόκλισης του αριστερού και του δεξιού αμφιωτικού ηχητικού σήματος, η οποία αποδεικνύεται ότι σχετίζεται με τα στατιστικά της αντίστοιχης μονοφωνικής ακουστικής απόκρισης. Τέλος, προτείνεται μια σειρά από παραμέτρους οι οποίες βασίζονται στα στατιστικά χαρακτηριστικά αμφιωτικών παραμέτρων και σχετίζονται με το αντιληπτικό μοντέλο της ανθρώπινης ακοής. Τέλος, προτείνεται μια νέα μέθοδος για την εκτίμηση της στάθμης λόγου κατευθείαν προς ανακλώμενου ήχου από στερεοφωνικά σήματα.

In speech communication systems such as voice-controlled systems, hands-free mobile telephones and hearing aids, the received signals are degraded by room reverberation and background noise. This degradation can reduce the perceived quality and intelligibility of the speech, and decrease the performance of speech enhancement and source localization. These problems are difficult to solve due to the colored and nonstationary nature of the speech signals, and features of the Room Impulse Response (RIR) such as its long duration and non-minimum phase. In this dissertation, we focus on two topics of speech enhancement and speaker localization in noisy reverberant environments. A two-stage speech enhancement method is presented to suppress both early and late reverberation in noisy speech using only one microphone. It is shown that this method works well even in highly reverberant rooms. Experiments under different acoustic conditions confirm that the proposed blind method is superior in terms of reducing early and late reverberation effects and noise compared to other well known single-microphone techniques in the literature. Time Difference Of Arrival (TDOA)-based methods usually provide the most accurate source localization in adverse conditions. The key issue for these methods is to accurately estimate the TDOA using the smallest number of microphones. Two robust Time Delay Estimation (TDE) methods are proposed which use the information from only two microphones. One method is based on adaptive inverse filtering which provides superior performance even in highly reverberant and moderately noisy conditions. It also has negligible failure estimation which makes it a reliable method in realistic environments. This method has high computational complexity due to the estimation in the first stage for the first microphone. As a result, it can not be applied in time-varying environments and real-time applications. Our second method improves this problem by introducing two effective preprocessing stages for the conventional Cross Correlation (CC)-based methods. The results obtained in different noisy reverberant conditions including a real and time-varying environment demonstrate that the proposed methods are superior compared to the conventional TDE methods. / Graduate / 0544 / 0984 / saeed.mosayyebpour@gmail.com