A substantial part of Acoustic Technology's research is carried out by PhD students, and most of our PhD students carry out their PhD projects in close cooperation with companies. Many of the projects are supported by the Industrial PhD Programme, funded by the Danish Agency for Science, Technology and Innovation.
Current PhD projects
Geok Lian Oh
Deconvolution of seismic transient signals: A model-based signal processing approach
The purpose of the planned PhD project is to examine and develop new ways of improving the deconvolution. The planned approach is model-based and will incorporate detailed knowledge of the physics of the signal. The project involves developing seismic wave field propagation models for transient sources and realistic topologies and geological conditions, and devising the model-based algorithms for the deconvolution task.
To be completed in 2014
Elisabet Tiana Roig
Acoustic array methods for identification of noise sources in vehicles
Signal processing techniques based on arrays of acoustic sensors for the localisation of sound sources are a powerful tool for the automotive industry, and the industry is currently incorporating systems of arrays of acoustic sensors during the design of vehicles for localising and identifying sources of noise. Such systems are used for measurements inside vehicles to characterise the noise sources and their effect on the comfort of the passengers. Besides, arrays are also used in measurements outside of vehicles for mapping environmental noise and for detection/prevention of structural failures. However, the complexity of these scenarios presents problems for the existing array techniques that need to be analysed and overcome.
Thus the purpose of this PhD study is to examine array techniques such as delay-and-sum, eigenbeamforming, adaptive beamforming, and deconvolution methods, identify their limitations, and establish an efficient methodology for integration of various methods in a single device.
To be completed in 2013
Nonlinear balanced armature receivers
Nonlinear distortion added by the loudspeaker (“receiver”) in a hearing aid reduces the signal-to-noise ratio in the acoustic output and may degrade the user’s ability to understand speech. The balanced-armature type loudspeakers predominantly used in hearing aids are inherently nonlinear devices, since any displacement of the loudspeaker diaphragm inevitably changes the magnetic and electrical characteristics of the loudspeaker. This project proposes to identify, model and control the key nonlinearities in such miniature balanced-armature loudspeakers. An accurate model of the nonlinear behaviour of these loudspeakers, will make it possible to design loudspeakers with less distortion and/or increased efficiency. This may be achieved by improved design and/or by active nonlinear compensation applied to the signal before it enters the loudspeaker.
The goal is to develop balanced-armature loudspeakers with increased sound quality which may ultimately result in increased speech intelligibility for the hearing aid user. An increase in loudspeaker efficiency is another important side-benefit since this may lead to extended battery lifetime.
The project is an industrial PhD study carried out in cooperation with Sonion.
To be completed in 2013
Antoni Torras Rosell
New measurement techniques: Optical methods for characterising sound fields
One of the fundamental problems in determining the spatial distribution of the quantities that describe a sound field is the influence of the transducer on the sound field itself when the transducer is immersed in the field. Typically, the influence of the transducer (in form of a frequency response) is assumed to be negligible when the size of the transducer is small compared with the wavelength of the sound wave or is rendered negligible by using a transducer-based correction that has been determined previously under well defined sound field condition, e.g., free field, uniform pressure, or diffuse field. Either solution introduces additional uncertainties to the measurement process. The use of optical techniques may help overcoming this problem because the sensing element is not a bulk transducer but a beam of light that does not influence the sound field at all, thus giving the actual value of the acoustic quantities.
The purpose of the project is to develop a set of methodologies for the application of optical, laser-based techniques to the measurement, characterisation and visualisation of sound fields, and to demonstrate that such measurements are possible with a precision similar to that of well-established classical measurement techniques.
The project is an industrial PhD study carried out in cooperation with Danish Fundamental Metrology.
To be completed in 2013
Completed PhD projects
Efrén Fernández Grande
Sound Source Identification with Acoustic Array Technology
Near field acoustic holography is a sound source identification technique that makes it possible to visualise the sound field in the near field of a source. It is particularly useful for studying complicated sound generation mechanisms. The technique involves measuring the sound field in a two-dimensional aperture near the source under test. Based on this measurement, the three-dimensional sound field near the source can be reconstructed. The resulting visualisation can provide valuable information about the source mechanism.
This project studies different techniques, their performance, limitations and applicability to different cases. It is of particular interest to investigate holography based on measurement of the particle velocity. This is due to the fact that kinetic energy dominates in the evanescent near field of the source. Moreover, it seems that the combination of sound pressure and particle velocity measurements can overcome some of the limitations of existing methods.
Completed in 2012
David Pelegrin Garcia
Speakers' comfort and increase of their voice level in lecture rooms
Vocal health is a major concern among teachers and represents a significant cost for the school authorities. The physical environment of the teaching room – not only the background noise – plays a very important role in determining the average voice power levels at which teachers speak. When designing the acoustics of a classroom, one has to predict quantitatively the vocal effort that the teachers will experience, and regulations regarding classroom acoustics should take account of the voice levels needed to speak in classrooms to improve the teacher’s vocal health.
In this project a virtual auditory environment is used for determining the voice levels that teachers use in different simulated classrooms under different background noise conditions. Additional measurements are carried out in real classrooms during teaching activities. Physical measures of the classroom and of the sound field are used to predict variations in the voice levels.
Completed in 2011
Modelling structural acoustic properties of loudspeakers
Modern loudspeaker cabinets are becoming increasingly complex. Consequently, the design engineers have to face the fact that unwanted structural acoustic problems may occur unexpectedly after a pre-production prototype has been manufactured and tested. Such problems are usually solved ad hoc at a late stage and may involve excessive costs because of the modification of production tools. Thus, prediction techniques are becoming more and more important. The objective of this project is to develop new techniques for predicting the structural acoustic properties and forced vibration response of loudspeaker cabinets of irregular geometry. More specifically, the goal is to develop a theoretical/numerical model that can simulate the mechanical and structural acoustic properties of a loudspeaker cabinet of any given geometry and predict the vibration in the low to mid frequency range.
The project is an industrial PhD study carried out in cooperation with Bang & Olufsen.
Completed in 2011
Unwanted feedback has been one of the major problems for users of hearing aids for a long time. Feedback occurs when a part of the amplified sound leaks from the ear canal and is picked up by the hearing aid microphone and then re-amplified. Unchecked feedback can lead to system instability limiting the maximum stable gain that can be achieved in hearing aids. The user experiences feedback as a very unpleasant loud whistling or howling sound.
The purpose of this project was to examine and develop new ways of improving feedback suppression techniques in hearing aids. The approaches included de-correlation techniques and new models for feedback paths. The project involved both computer simulations and laboratory experiments. A breakthrough in feedback suppression will bring better comfort to the hearing-aid users and improve the design of the current hearing aids.
The project was an industrial PhD study carried out in cooperation with ReSound.
Completed in 2010
Investigation of internal feedback in hearing aids
Modern hearing aids are met with demanding aesthetic requirements such as minimal physical size and visibility. The loudspeaker and the microphones are thus placed closely together. Consequently, problems with the transmission of sound and vibrations from the loudspeaker to the microphones easily occur. This causes a feedback problem, and this limits the obtainable amplification and therefore represents a critical design problem.
The feedback problem has been investigated by developing a full 3D-model of a behind-the-ear hearing aid. The model simulated the vibroacoustic transmission from the loudspeaker to the microphones. The hearing aid skeleton was based on finite element analysis, whereas the internals were modelled using the theory of “fuzzy structures”. Simulations showed a close agreement with electroacoustic measurements on the actual hearing aid.
The project was an industrial PhD study carried out in cooperation with Widex.
Completed in 2009
Active noise cancellation headsets
Extremely high sound pressure levels are encountered onboard airborne military platforms, and even though the pilots wear headsets with significant passive noise attenuation they are nevertheless exposed to very high sound pressure levels that affect their situational awareness. A possible solution could be hearing protectors with active noise control. The purpose of the project was to examine various active noise control strategies. Active noise reduction systems based on feedback control are limited by time delays; and active noise reduction systems based on feedforward control are limited by lack of coherence between the reference signals and the error signals to be minimised; therefore a considerable part of the research has been focused on examining these limitations.
The solution was a control system that combined feedforward and feedback control and used adaptive algorithms that can handle the impulsive signals occurring in military airplanes and helicopters.
The project was carried out in cooperation with Terma.
Completed in 2008
Ruiz Villamil, Heidi (former guest PhD)
Optimization of Sound Absorbing Systems Based on Multiple Micro-Perforated Panels