Analysis and morphing of monaural cues from individual head-related transfer functions

Master Thesis of Braren, Hark
Betreuer: Fels, Janina / Bomhardt, Ramona

The core of every 3D audio presentation is the head related transfer function (HRTF). It is quite specific to each individual, due to varying expressions of head shape and outer ear (pinna) geometries among humans. Even though some applications work with general HRTFs (i.e. from a dummy head), it is absolutely necessary to use individual transfer functions in order to achieve a truly immersive experience with correct positioning of virtual sources in the 3D space. Measuring the HRTFs takes specialized equipment and a lot of time. The goal of this thesis is to develop a method that can generate individualized head related transfer functions from a set of head and pinna measurements. The proposed method uses principal component analysis on a database of known HRTFs. Multiple regression analysis is used to find a direction specific relation between the principal components and the pinna and head dimensions. These relations can then be used to calculate an individualized set of HRTFs for an unknown subject from their personal pinna and head dimensions. In this thesis several methods to improve the accuracy of the individualized HRTFs are presented and evaluated. Additionally, the problem of inaccuracies in the measurement of pinna dimensions is discussed and the foundation of an automated measurement process using scans or photographs of the ear is presented.

Assessment of soundscapes through different approaches: soundwalks and listening tests

Master Thesis of Su, Yuanyuan
Forschungsgebiet: Lärmforschung / Psychoakustik
Betreuer: Engel, Margret / Fels, Janina

Soundscape is related to the relationships between the ear, human beings, sound environments, and society. Based on a series of case studies in Europe and China and an intensive literature review, the description, evaluation, and built for soundscape in urban open spaces are systematically analyzed in terms of four basic elements: sound, space, people, and environment. For urban open public spaces, it is critical to study the peoples’ perception of sounds. This research aims to establish the relation between the auditory perceptions and the visual aspects. First of all, what is the influence of the visual aspects on the perception of the urban soundscape? This question will be studied based on the sample recording and a large number of soundwalks in two open public locations in Aachen. The recordings will be performed by using binaural and omni-directional microphones. Another method to collect the data is listening test. The sample recordings with the same participants from soundwalks will be used and calibrated. Additionally the subjective thresholds from psychoacoustic parameters in percentiles, including loudness, sharpness, fluctuation strength, roughness and tonality will be established. Therefore the psychoacoustic parameters for the soundscape recordings will be calculated by using the software ArtemiS. At last, which is the best method to collect data for soundscape studies will be determined by neural network. It will also be identified if the statistical method is suitable for subjective and objective data fitting.

Determination of listening effort in primary-school aged children using a dual-task paradigm with different noise situations.

Bachelor Thesis of Burger, Julian
Forschungsgebiet: Auditorische Szenenanalyse / Lärmforschung
Betreuer: Loh, Karin

Listening effort has not been intensively researched in primary-school aged (6-10 years) children yet. It refers to the attention and the cognitive resources needed to understand speech. A child-oriented dual-task paradigm is developed in order to determine the children's listening effort in different noise situations. The experiment will be conducted in a mobile hearing booth with 24 normal-hearing children. The primary task is to identify a word presented in noise by hitting the button with the matching picture out of four pictures. The secondary task includes reacting to a flashing light on screen, also, by hitting a button. It is hypothesized that children require more cognitive resources in a noise condition than in a silent condition when trying to understand speech.

Development and Evaluation of an FPGA based Illtrasoiind Hardware for Prosthodontics

Master Thesis of Kotzur, Benjamin
Forschungsgebiet: Sonstige
Betreuer: Fels, Janina

Der Prozess zur Herstellung von Zahnrestaurationen erfordert eine Vielzahl aufwendiger Schritte, die konventionell mit der analogen Abformung im Patientenmund beginnen. Aufgrund der Fehleranfälligkeit, mangelnden Effizienz gegenüber CAD/CAM basierten Methoden sowie der fehlenden Möglichkeit, unter dem Zahnfleisch liegende Präparatiornsränder abzuformen, wird am Lehrstuhl für Medizintechnik ein digitales, auf Ultraschall basierendes, Verfahren zur intraoralen Datenabnahme präparierter Zähne entwickelt. Das aktuellle Labormuster des intraoralen Ultraschallscanners basiert auf einem mechanisch bewegten Einzelelementwandler. In vorangegangenen Arbeiten wurde ein miniaturisiertes und kostengünstiges Ansteuerungssystem auf Basis eines digitalen Signalprozessors (DSP) entwickelt. Im Zuge der Entwicklung eines Prototypens soll nun unter Berücksichtigung der hohen Sicherheits- und Genauigkeitsanforderungen ein Ultraschallmodul bestehend aus Pulser und Verstärker entwickelt, in das bestehende System integriert und getestet werden. Ein weiterer Schwerpunkt liegt auf der Gewährleistung einer hohen Datenübertragungsrate. Nachdem der medizinische und technische Hintergrund aufgearbeitet wurde, soll im Rahmen dieser Masterarbeit zunächst ein geeignetes Konzept für die gerätetechnische Umsetzung erarbeitet werden. Die hierfür notwendigen Software- und Hardware-Komponenten inkl. der Signalverarbeitung zur Messdatenauswertung müssen implementiert werden. Hierzu sollen am Lehrstuhl vorhandene Komponenten der Ultraschallhardware sowie der Steuerungs- und Regelungshardware für diese Anwendung weiterentwickelt und optimiert werden. Anschließend soll ein Versuchsstand konzipiert, aufgebaut und in Betrieb genommen werden. An diesem Versuchsstand sollen abschließend Ultraschallmessungen zur Evaluierung des entwickelten Ultraschallsystems durchgeführt werden. Die Vorgehensweise und Ergebnisse der Arbeit sollen klar und umfassend dokumentiert und diskutiert werden.

Development and Evaluation of an Ultrasound System for Bone Quality Assessment in Dental Implantology

Master Thesis of Grüterich, Martin
Betreuer: Fels, Janina

The aim of the present study was the development of a concept for application of ultrasound (US) teclmology in preoperative assessment and evaluation of the available bone quality for dental implantology. This is believed to be a crucial requirement for optimal implant planning and makes an important foundation for choosing a suitable implant system. TIvo system components were developed: The first component utilizes a variable low-frequency (NF) US-frequency about 1 MHz. The second one uses a high frequency (HF) 3D-US-miclVscanner at 45 MHz. For concept development and evaluating the NF-system a simulation environment based on MATLAB and PZFlex was developed. Afterwards Polymethylmethacrylat(PMMA) samples of varying thickness were produced and measured as models for human cortical jawbone. It was demonstrated that the amplitude of the reflected US signals changed due to intelference phenomenons with the thickness of the PMMA samples. Based upon these local minima an algorithm was developed which is able to reconstruct the sample thickness out of a predefined measuring procedure. The HF-system was tested on human skull bone assembled with dental implants. Additional a series of measurements was performed on jawbone models made from porcine ribs. Possible deviations for the NF-system were calculated to ±5.66%. For the HF-system the absolut accuracy is 13,8 flm with a precision of 16.9flm. The bone defects located in the jawbone models could be measured with a maximal deviation of 0.96 mm

Development of a demonstration and evaluation tool for individualized binaural technology

Master Thesis of Doma, Shaimaa
Forschungsgebiet: Binauraltechnik
Betreuer: Braren, Hark / Fels, Janina

The transfer function from an arbitrary source direction to the ear canal entrance is characterized by typical interactions with head, pinna and torso geometry. This highly individual filtering of sound is described by the head-related transfer function (HRTF) and is essential for a realistic spatial auditory reproduction. Obtaining individual HRTFs of high quality, however, requires specialized equipment that is not widely available. In some cases, it has shown to be particularly difficult, e.g. for children, whose movement during the measurement leads to inaccurate results. Individualized HRTFs offer a compromise by making modifications to a given non-individual HRTF set. Over the past years, various approaches have been discussed, such as the individualization of the Interaural Time Difference (ITD), frequency scaling or methods based on Principal Component Analysis (PCA). These methods use anthropometric dimensions to estimate a subject’s individual HRTF. The dimensions can, for instance, be obtained from 3D images generated by structured light scanners or magnetic resonance imaging (MRI). In this thesis, a tool is to be developed which allows for a direct auralization whilst manipulating the parameters of several of the aforementioned HRTF individualization methods. Using a set of exemplary virtual acoustic scenes, the user shall be able to modify a provided HRTF set. Applications of the tool comprise the subjective optimization and evaluation of HRTFs.

Development of an algorithm for automated segmentation of tooth surface in 3D Ultrasound data of subgingival prepared teeth based on machine learning

Master Thesis of Mündlein, Fabian
Betreuer: Fels, Janina

Estimation of Direction Dependent Pinna Reflection Locations

Bachelor Thesis of Biot, Jérome
Forschungsgebiet: Binauraltechnik
Betreuer: Braren, Hark

Head-related transfer functions (HRTFs) describe the direction dependent properties of the human hearing apparatus. Apart from head and and torso shapes, it is primarily the individual geometries of outer ear, or pinna, that play an important role in the spectral shaping, which the brain in turn uses as direction specific cues. By using a high-resolution 3D Scanner (like the Artec Space Spider), we can accuratly measure and digitize these geometries. As part of the thesis, this scanner is to be used to measure a subjects outer ear and the determine the direction specific locations of the first order reflections. This information is to be used to estimate the spectral coloration at the ear canal, which in turn can be validated by measuring the subjects HRTF.

Evaluation of the effect of continuous rotation during measurements of head-related transfer functions

Bachelor Thesis of Shaima’a Ahmed Bassim Ali Doma
Forschungsgebiet: Binauraltechnik
Betreuer: Richter, Jan-Gerrit / Fels, Janina

Hearing aid signal processing for a real-time auralization system

Master Thesis of Tumbrägel, Jonas
Betreuer: Pausch, Florian / Fels, Janina

Signal processing algorithms for hearing aids have been extensively investigated and improved in the last decade and are usually adjusted under laboratory conditions. In real-world situations however, hearing aid users still report an unsatisfying sound perception, especially within noisy environments. In order to increase the performance of hearing aids in difficult everyday situations, hearing aids algorithms need to be evaluated in more realistic test scenarios in cooperation with HA users. Suitable test conditions in real-world environments cannot be created efficiently or lack in flexibility. Virtual acoustic environments (VAE) are used to simulate real-world acoustic scenes and to enable controlled listening conditions. To avoid inaccurate sound field reproduction for hearing aids in VAE, simulated signals can be processed with hearing aid algorithms and subsequently inserted to a dummy hearing aid (DHA). Hearing aid users with mild to moderate hearing loss do not only perceive the surrounding sound field amplified through the hearing aid, but still have remaining hearing capabilities which have to be addressed. This is done by expanding the system by a loudspeaker-based reproduction in combination with a crosstalk cancellation filter network. In this work, different hearing aid signal processing algorithms implemented on a Master Hearing Aid platform are investigated and evaluated. Additionally, measurements of the electroacoustic properties of a DHA have been carried out. Finally, these two system components have been integrated into the VAE and were tested in a combined evaluation.

Implementation and Evaluation of a Trust Region Algorithm for Optimization of Multiway Crossover Filters under Constraints

Bachelor Thesis of Henzel, Robert
Betreuer: Richter, Jan-Gerrit / Fels, Janina

Influence of continuous subject rotation on HRTF measurement data

Forschungsgebiet: Binauraltechnik / Akustische Virtuelle Realität
Betreuer: Richter, Jan-Gerrit

The measurement duration of head-related transfer function measurements has been greatly reduced in recent years. A shorter measurement time will reduce strain on the subject as well as reduce movement which otherwise introduces errors in the measurement. To further reduce the measurement time, a continuous rotation approach was investigated objectivly. A subjective evaluation of the measured data is needed.

Influence of Continuously Tracked Head Movements on Intentional Switching in Auditory Selective Attention

Bachelor Thesis of Wepner, Saskia
Forschungsgebiet: Auditorische Szenenanalyse
Betreuer: Oberem, Josefa

Bisher erfolgten Hörversuche zum intentionalen Aufmerksamkeitswechsel mit statischer Wiedergabe der Ziel- und Störsprecher. Vergangene Experimente geben jedoch Grund zur Annahme, dass eine realitätsnähere Darbietung von auditiven Szenen die kognitive Leistung von Versuchspersonen verbessern kann. So ist bereits bei der Erweiterung von Experimenten mit dichotischer auf eine binaurale Wiedergabe der akustischen Reize erkennbar, dass die kognitive Leistung der Probanden deutlich gesteigert werden kann -- dies schlägt sich in verminderten Reaktionszeiten sowie niedrigeren Fehlerraten nieder. Zur (besseren) Lokalisation(/räumlichen Wahrnehmung) von Schallquellen in auditiven Szenen stellen individuelle Peilbewegungen des Kopfes eine weitere wichtige Funktion dar, da diese eine schnellere und korrektere Einordnung der dargebotenen akustischen Reize ermöglichen. Diese Arbeit soll Aufschluss über den Gewinn der Erstellung eines realistischeren Hörerlebnisses durch die entsprechende Erweiterung von auditiven Szenen geben. Um die akustische Szene weiter an reale Bedingungen anzupassen, sollen Kopfbewegungen von Probanden mit in Versuche zur auditiven selektiven Aufmerksamkeit (ASA) einbezogen werden. Teil der Arbeit sind daher die Implementierung sowie Durchführung eines Hörversuchs zur Überprüfung \uwave{dieses Paradigmas}. Zur Detektion der Peilbewegungen soll ein visueller Tracker verwendet werden mit dessen Hilfe die akustischen Reize an die Bewegungen des Probanden angepasst werden können. Es ist zu erwarten, dass sich die Lokalisationsleistung der Probanden in Hörversuchen durch Hinzunahme individueller Peilbewegungen weiter verbessern wird.

Influence of destructive interferences of the head-related transfer functions on localization

Bachelor Thesis of Ruhnke, Wojciech
Betreuer: Fels, Janina / Bomhardt, Ramona

Influence of fine structure in head-related transfer function data

Forschungsgebiet: Binauraltechnik / Akustische Virtuelle Realität
Betreuer: Richter, Jan-Gerrit

For virtual room acoustics simulations a good binaural synthesis is needed. This allows virtual sound sources to be placed at arbitrary directions. This is achieved using head-related transfer functions (HRTFs). These functions can be described as a frequency dependent directivity with fine spatial structure. This work should focus on noticeable differences of this fine spatial structure of head-related transfer function data. It is assumed that only a certain amount of detail is needed for most binaural synthesis application. An investigation into a just noticeable difference will give an upper limit on the amount of detail. To this end, a listening test needs to be planned and executed

Influence of noise on listening effort of adults

Bachelor Thesis of Batlle, Roser
Forschungsgebiet: Auditorische Szenenanalyse / Lärmforschung
Betreuer: Loh, Karin

This thesis consists of a study with adults in respect to educational institutions with the aim to evaluate speech perception and listening effort depending on the noise condition. For this purpose three different noise conditions will be simulated: speech-shaped noise (SSN), which is characterized for having a similar spectrum as the average long-term spectrum of a speech signal and is considered to be the most effective energy masker of speech; irrelevant and relevant multi-talker babble (IMTB and RMTB, respectively), which consists of several people simultaneously talking in a foreign language for the listener, in IMTB, or in the same or similar language, in RMTB. These noises will be included individually as background noises in a dual-task listening experiment. The first task will consist of the measurement of time response and percentage of correct words, where participants will have to identify stimuli word and select the corresponding image between four possible options (e.g. Kopf - Topf - Zopf - Frosch). Then, the second task will focus on the measurement of reaction time after a visual stimuli, in which participants have to press a button when the light bulb in the screen turns on. This experiment will test a total of 24 conditions by modifying signal-to-noise ratio, target and distractor position combination and type of noise.

Influence of subject movement on individual measurements of head-related transfer functions

Master Thesis of Wepner, Saskia
Forschungsgebiet: Binauraltechnik
Betreuer: Fels, Janina / Richter, Jan-Gerrit

In recent years, the need for individually measured head-related transfer functions (HRTFs) has increased steadily. For instance, HRTFs are important for the development of psychological experiments including spatial perception or multi-speaker scenery; as well as in industrial fields, such as virtual reality scenes. Therefore, measurements need to be sped up and simplified. Current research offers various approaches to develop less time-consuming alternatives to prior measurement methods. As errors may influence the auditory impression of reproduced signals, best possible conditions are demanded as well as a detailed knowledge of uncertainties occurring during the measurement. Certainly, additional inaccuracies do appear due to natural movements of the subject. The aim of this thesis is to explore how and to what extent subjects move during the measurement procedure. Hence, it is examined which measurement method does evoke least movements of test subjects. Additionally, a real-time feedback system will be installed, that allows the subject to perform a correction of position and orientation. The gained knowledge shall also reveal the less error-prone measurement method and allow for a better reconstruction of measurement signals.

Influence of the artificial torso of the dummy head

Forschungsgebiet: Binauraltechnik / Psychoakustik
Betreuer: Richter, Jan-Gerrit

The localization ability of humans stem from the influence by the head and pinna shape on the sound field. To recreate such an influence, it is often more efficient to have a fixed head model. These models, or artificial heads give a good but not individually perfect spatial impression of sound sources. This thesis focuses on the influence of the torso of an artificial head that is used both as a reference measurement device and an important tool for many virtual acoustics scenarios. It is studied with the use of Boundary Element Method simulation which will allow an easy way to examine different types of material as well as different geometries. The goal of the thesis is to create a simulation that is similar to measurements with a wooden torso while gaining a better understanding of the importance of the torso to the head.

Influence on the localization of virtual sound sources using head-related transfer functions with smoothed phase spectra

Bachelor Thesis of Vollmer, Lukas
Betreuer: Bomhardt, Ramona / Pausch, Florian

Head-related transfer-functions (HRTF) describe the influence of body and head on the sound that arrives at the earcanal. This sound contains directional information, which can be extracted by the human ear. One of these cues is the interaural time difference (ITD) that quantifies the time delay between the arrivals of the sound at the left and right ear, respectively. In the frequency domian, time delays are representet as a linear phase but the phase spectra of HRTFs are only approximatly linear. To linearise a function, smoothing algorithms can be used. In terms of HRTFs, these algorithms were mostly used to smooth the magnitude, while the phase spectrum was maintained. Hence, in this study the influence of smoothed phase spectra on the localisation of virtual sound sources is investigated. Therefore different smoothing algorithms are presented and analysed. Because of the smoothing, errors in the ITD are expected, but the analysis shows that phonetic changes occur before the errors in the ITD become perceptible. Thus, instead of a localisation experiment, a perceptional listening experiment was conducted. The results show that the influence of the smoothing, compared to the objective results, has to be reduced, so that no difference between smoothed and original phase is perceptible.

Investigation effects of auditory training under realistic acoustics in virtual reality

Master Thesis of Ray, Alokeparna

A person’s ability to comprehend speech is vital for communication in social, educational, professional and cultural contexts. However, in most surroundings speech is masked by noises, which can distract listeners in comprehending speeches from the target speaker. In binaural hearing in adverse acoustics, spatial cues from the sound sources are important for speech recognition, by providing resources for listeners to distinguish target talker from distracting babble noises. The aim of this thesis is to develop a paradigm using realistic acoustic virtual reality to train individuals in using spatial cues under adverse acoustics. The training is conducted in a real-time virtual acoustic environment emulating a typical classroom learning scenario with a spatial release from masking task, in which spatial cues will be provided for speech recognition. Participants are trained in either anechoic or reverberant condition. To examine the training effect, assessment on speech reception thresholds are administered both before and after training.

Investigation of binarual reproduction with headphones for biomedical imaging applications

Master Thesis of Nguyen, Quoc-Viet
Betreuer: Fels, Janina / Richter, Jan-Gerrit

The binaural reproduction of sound signals has been studied vastly with regards towards plausibility and authenticity. In terms of the perception inside the brain, however, research has just recently started as brain analysis via fMRT needs high magnetic fields in order to track the region and magnitude of brain stimulation. Generating and sustaining these magnetic field creates unavoidable loud noises. Due to these constraints, audio systems used in such tests have to provide sufficient electromagnetic and acoustic shielding. Therefore, a trade-off between binaural reproduction quality and shielding is taking place in the development of these systems. The aim of this thesis is to analyse a proprietary headphone system with regards to its spatial reproduction capabilities and to conduct improvements on the headphone system with regards to binaural reproduction. To this end, two listening experiments are condncted. After the first experiment, the headphones were modified from a closed towards an open system. Apart from improvement in localisation properties, the modification was expected to change the frequency response of the headphones. The conducted experiments focused on the comparison of different binaural stimuli and were carried out by two independent test groups. All in all, 23 untrained subjects participated in the first experiment round with the unmodified headphones and were later compared to the performance of the 33 subjects in the second round. Additional measurements regarding the headphones' characteristics have been taken as well. The results showed a significant improvement in all tested spatial conditions.

Investigation of noise exposure in educational institutions using aurally-accurate techniques for preschool and primary school-aged children

Master Thesis of Loh, Karin
Forschungsgebiet: Psychoakustik / Lärmforschung
Betreuer: Fels, Janina / Pausch, Florian

Investigation of realistic acoustic scene simulation for people with hearing impairment

Master Thesis of Stockmann, Anne
Betreuer: Pausch, Florian / Peng, Zhao

The simulation of room acoustic environments (Virtual Acoustic Environments, VAE) has potential for a variety of useful medical applications, such as the evaluation of new hearing aids and hearing aid algorithms. To further the use of VAE into the clinical population, this thesis first focuses on the room acoustic and psychoacoustic parameters in classrooms from in situ measurements, which provide input on the simulation of realistic acoustics in future studies for children. The room acoustics measurements took place in unoccupied classrooms at a school for children with educational needs in hearing and communications. Activities during lessons and lunch breaks were also recorded and analyzed. The second part of the present work evaluates a loudspeaker-based binaural playback system with research hearing aids through normal hearing adults’ performance of localization of virtual sound sources. The data is compared to the results of a previous localization study using virtual sound sources played via headphones and real sound sources.

Investigation of the influence of small head movements on the localization ability of human listeners

Master Thesis of Wang, Suliang
Betreuer: Fels, Janina / Richter, Jan-Gerrit

Investigations of binaural speech transmission index in classroom acoustics

Forschungsgebiet: Psychoakustik / Binauraltechnik
Betreuer: Loh, Karin

Lärmbelastungen in Bildungseinrichtungen werden heutzutage mit konventionellen und standardisierten Messmethoden erfasst und bewertet. Die dazu ermittelten Einzahlkennwerte, wie der A-bewertete Schalldruckpegel, können jedoch die Komplexität des Lärms und dessen Wahrnehmung nicht richtig wiederspiegeln. In der Psychoakustik und in der Raumakustik existieren bereits erste (binaurale) Modelle, die die Lärmwahrnehmung in geschlossenen Räumen genauer beschreiben sollen. Diese Modelle jedoch wurden mittels Forschungsergebnissen entwickelt, in denen Erwachsene die Basis der Untersuchungen darstellten. In Bezug auf Bildungseinrichtungen ist jedoch noch unbekannt, ob diese Modelle so auch für Kinder gültig sind, da diese nicht nur eine andere Körpergeometrie, sondern auch eine andere kognitive Wahrnehmung ihrer Umgebung haben. In dieser Arbeit sollen die Auswirkungen von verschieden Erfassungsmethoden des binauralen Sprachübertragungsindexes untersucht und mittels eines Hörexperiments ausgewertet werden.

Investigations on head movements in localization experiments with different dynamic binaural replay methods

Master Thesis of Setzer, Dorothea
Forschungsgebiet: Auditorische Szenenanalyse / Binauraltechnik
Betreuer: Oberem, Josefa / Richter, Jan-Gerrit

In the context of binaural-listening paradigms to be rendered more realistic, previous studies have examined the influence of head-movements on the human capacity of sound localization. This thesis enhances static binaural reproduction by allowing for headmovements and with individually measured HRTF data and real-time head tracking permits a dynamic binaural reproduction of sound sources via headphones. Sound localization is evaluated for 20 subjects who were presented a set of sound sources at crucial positions as on and near to the median plane and on a Cone of Confusion. Head-movements, as well as error rate and reaction time are registered comparing the results for the individually measured HRTFs with a resolution of (5° x1°) , the non-individual HRTFs of (5° x1°) with and without interpolation each and a set of non-individual HRTFs of (1°x1°) in azimuth and elevation. The recently introduced technique of proximal pointing is used in order to obtain accurate results for pointing angles in the localization task.

Investigations on the just noticeable difference of wall reflexions regarding the wall position and orientation in virtual rooms

Bachelor Thesis of Schäfer, Johanna

Klanglandschaft Stadt. Einfluss von Geräuschen des städtischen Umfeldes auf die Lebensqualität in drei Aachener Quartieren

Bachelor Thesis of Carls, Michele
Forschungsgebiet: Lärmforschung / Psychoakustik
Betreuer: Engel, Margret

Localization accuracy comparing real and virtual sources

Forschungsgebiet: Binauraltechnik / Psychoakustik
Betreuer: Richter, Jan-Gerrit

The localization ability of humans stem from the influence by the head and pinna shape on the sound field. This influence can be measured and is called Head related transfer function (HRTF). HRTFs can be used to create virtual sound sources at arbitrary places. This principle is heavily used in virtual acoustics and room acoustics simulations. It was shown that the ability to localization sound sources is dependent on the used HRTF. Generally, real sound sources are localized with the highest accuracy. This thesis focuses on a comparison of localization accuracy. It is used as a measure on performance of the whole binaural synthesis application developed at the institute. It will make use of all the state of the art measurement and reproduction equipment to achieve virtual sound sources with the highest localization accuracy. As a comparison, real sound sources are used and a listening experiment is to be conducted. The goal is to quantify the quality of binaural synthesis and virtual acoustics.

Matching of objective psychoacoustic parameters and subjective perception of soundscapes in Aachen

Master Thesis of Sethi, Anuj
Betreuer: Fels, Janina / Engel, Margret

Methods for the evaluation of rotational-speed dependent noises with spectro-temporal patterns

Master Thesis of Becker, Henri
Betreuer: Fels, Janina

Rotational-speed dependent noises with modulation and prominent temporal stmctures are often perceived as annoying. If the sound of a technical product exhibits these characteristics, it is most likely associated with poor quality. In the scope of this master thesis the typical characteristics of such noise phenomena will be investigated in time and ti'equency domain using established methods (e. g., Modulation Spectrum, Relative Approach). Noise variants with differently pronounced patterns and modulation will be generated by systematically manipulating the source. These sounds will be evaluated in listening tests with regard to a quality criterion. FurthelIDore, based on the results of the listening tests, analysis methods shall be presented as the basis for an instrumental assessment of the interesting noise phenomena. ,-

Objective and subjective analysis of room acoustics in open-plan offices

Master Thesis of Xu, Xinya
Forschungsgebiet: Psychoakustik / Lärmforschung
Betreuer: Loh, Karin

In day-to-day work, people are surrounded by various noises that can significantly affect cognitive performance and acoustic satisfaction. This acoustic perception and its effect are caused and influenced by various factors. Therefore, a diverse acoustic spatial analysis, from the objective as well as the subjective aspect, is needed. In this thesis, open-plan offices are measured with aurally-accurate methods and their acoustic properties are analyzed. On the one hand the room acoustics of the rooms are examined, on the other hand the background noise in the presence of persons is measured for the psychoacoustic analysis (in-situ measurements). The ITA artificial head is used to get an aurally-accurate evaluation. In addition, the subjective perception is examined using questionnaires, which are filled out by the persons present in the rooms during the in situ measurements. Subsequently, relations between room acoustics and psychoacoustic parameters, as well as between objective and subjective results are evaluated.

Perception and Modeling of sharpness of stationary and time-varying sounds

Master Thesis of Tarasova, Valeria
Betreuer: Fels, Janina

This master thesis deals with the investigation of various factors that influence the perception of the psychoacoustic parameter sharpness both for stationary and time-varying sounds. Present calculation methods of sharpness consider the "center of gravity"' of the weighted specific loudness patterns as a measure for sharpness. Thus their results do not depend on loudness.Only the Aures model includes, to a small extent. the influence of loudness on sharpness by an additional loudnessdependent factor.Until now mostly stationary synthetic sounds (filtered noise and pure tones) haw been used for the derivation of sharpness models. However, when these calculation procedures are applied to generally time-varying technical sounds. The results deviate significantly' from the listening test results.This study new listening tests performed to investigate systematically the influence of loudness as well as temporal structures on sharpness. The findings from the listening tests should serve as a basis for an imprm'ed model of sharpness perception for stationary and time-varying sounds.

Phasenrekonstruktion und Gewichtung für gemittelte Kopfübertragungsfunktionen

Bachelor Thesis of Bechtel, Phillip
Betreuer: Bomhardt, Ramona

To present people a spatial sound through headphones, it is necessary to determine the head related transfer function HRTF, for each desired position of the virtual sound source. These HRTF include information on the position of sound source, as well as the geometrical dimensions of the perSOll, for example shoulder, head and ear dimensions. As it is too expensive to carry out individual measurements for each person, an average, human head, a so-called dummy head was developed. With this dummy head an HRTF-Set is measured and fitted. Fitting is the approximation of a transfer function with poles and zeros. Fitting makes a HRTF -Set compact and reduces the memory overhead. The aim of fitting is to generate the best possible approximation of the measured HRTF so that no characteristics get lost. Shifting the poles and zeros of a fitted HRTF-Set, set the changes. This shifting allows individualization of a HRTF-Set. Consequently the number of poles and zeros should be kept as low as possible. Otherwise too many free parameters have to be determined for a potential individualization. In this paper an algorithm has been developed which divides the considered frequency band in smaller areas. These areas are fitted and joined with band pass back together. This results in smaller, independent solution vectors. To get the same result with a conventional fitting algorithm, the same number of poles and zeros are necessary. Generally the number of required poles and zeros depends on the desired quality of the fit and on the resolution of the measured HRTF. To evaluate the developed algorithm the case with a divisioll into four frequency bands was compared with a conventional algorithm. For comparison, the interaural time difference (ITD), interaural level difference (ILD), monaural cnes and the spectral difference were considered. The results of this comparison show that both algorithms provide similar results with an identical number of poles and zeros. The fitted HRTFs of the conventional algorithm is a bit better in the lower frequency ranges and the newly developed algorithm performs better at higher frequencies. Thus the phase is reconstructed more accurately at higher frequencies.

Sensitivity to changes in masker velocity in an Intelligibility task with virtual acoustics

Forschungsgebiet: Auditorische Szenenanalyse / Psychoakustik
Betreuer: Viveros, Rhoddy

Assess the main intelligibility indexes (Speech Reception Threshold (SRT) and Spatial Release from Masking (SRM)) for different masker(s) velocities in a listening-in-noise task and analyze how affect the comprehension capacity (how change SRT and SRM) when the masker(s) moves at different velocities. Listening test(s) is(are) necessary and Matlab knowledge. This theses can also be a mastertheses

Spatial advantages in intelligibility for lineal moving masker(s) and listener self head movement with virtual acoustics

Forschungsgebiet: Auditorische Szenenanalyse / Psychoakustik
Betreuer: Viveros, Rhoddy

Assessment of the Intelligibility advantages in a listening-in-noise task under a lineal moving masker(s) paradigm. Is necessary the evaluation of the indexes “Speech Reception Threshold” and “Spatial Release from Masking” to assess how change the intelligibility under this paradigm, taken into account the complexity of a Signal to Noise Ratio not constant, changing each instant during the test. Moreover the listeners self head movements for each condition in the listening test need to be measure to know if there are benefits in the comprehension capacity because the free (natural) movements of the head. Listening test(s) is(are) necessary and Matlab knowledge.

Subjektiver Vergleich von gemessenen und individualisierten Außenohrübertragungsfunktionen bei binauraler Wiedergabe mit Attributen zur Beschreibung der Wahrnehmungsqualität

Bachelor Thesis of Heiser, Ben
Forschungsgebiet: Binauraltechnik / Psychoakustik
Betreuer: Richter, Jan-Gerrit

In order to improve binaural reproduction, individual head-related trans- fer functions (HRTFs) are used. In this work, two existing methods of ac- quiring individual HRTFs are compared. Both methods are based on differ- ent strategies and require different hardware. HRTFs are measured using a loudspeaker array with interleaved sweeps as excitation. The second method applies individualization strategies. It consists of applying a principal com- ponent analysis on the data set from a three dimensional ear scan under consideration of anthropometric dimensions. A subjective comparison of the binaural reproduction will be done by using verbal descriptors of perceptual qualities. A listening experiment using dynamic binaural reproduction will be conducted to compare the perceptual qualities of both methods with real sources as reference.

Validation of an ellipsoidal model for the individualization of arbitrary Head-Related Impulse Responses

Master Thesis of Parvizian, Farzaneh
Forschungsgebiet: Binauraltechnik
Betreuer: Bomhardt, Ramona

In this thesis different subjective evaluation method for the analytical ellipsoid model of Lins & Bomhardt are evaluated. With this model it is possible to adapt the interaural time difference (ITD) of a given HRTF to an arbitrary subject. Therefore different types for the determination of an ITD from an HRTF are presented. Then two different kinds of listening test are designed for the evaluation. It turns out that the first method doesn’t work well for the evaluation of the model. Here the subject two sounds with different ITDs are presented to the subject. Then the subject was asked which of them are closer to a given direction which is marked in the listening room with a light. Because of the disadvantages of this kind of pointing method it was tested which subjects can hear a difference between the original ITD and the adapted ITD. Thus, for the model the head width was measured by a caliper and in addition the ITD of the subject is determined an HRTF measurement in front of the listening test. For the test itself a 3AFC method was used to compare once the original and the adapted ITD from caliper as well as the original and the adapted ITD from the HRTF. It figured out that most of the subject’s whose head is 1.5cm smaller than the original can distinguish between both ITDs. In addition it can be found that with increasing difference in the head width the performance to distinguish between both increases. The comparison between the ITD from the caliper and the subjects’ HRTF it shows that it is harder to detect the difference between the original ITD and one from the measured HRTF of a subject. In this thesis different subjective evaluation method for the analytical ellipsoid model of Lins & Bomhardt are evaluated. With this model it is possible to adapt the interaural time difference (ITD) of a given HRTF to an arbitrary subject. Therefore different types for the determination of an ITD from an HRTF are presented. Then two different kinds of listening test are designed for the evaluation. It turns out that the first method doesn’t work well for the evaluation of the model. Here the subject two sounds with different ITDs are presented to the subject. Then the subject was asked which of them are closer to a given direction which is marked in the listening room with a light. Because of the disadvantages of this kind of pointing method it was tested which subjects can hear a difference between the original ITD and the adapted ITD. Thus, for the model the head width was measured by a caliper and in addition the ITD of the subject is determined an HRTF measurement in front of the listening test. For the test itself a 3AFC method was used to compare once the original and the adapted ITD from caliper as well as the original and the adapted ITD from the HRTF. It figured out that most of the subject’s whose head is 1.5cm smaller than the original can distinguish between both ITDs. In addition it can be found that with increasing difference in the head width the performance to distinguish between both increases. The comparison between the ITD from the caliper and the subjects’ HRTF it shows that it is harder to detect the difference between the original ITD and one from the measured HRTF of a subject.