PhD Studentship: The diffuse sound object



Ref: ISVR-FDAG-114

Deadline: 1st September 2013

Open PhD position at the Institute of Sound and Vibration Research at the University of Southampton, in collaboration with BBC Research & Development

Summary of student eligibility criteria:

• First-class honours MSc or equivalent degree in Engineering, Physics or Mathematics

• Strong background in signal processing, acoustics and audio technologies

• Ordinarily resident in the UK for 3 years prior to the start of the grant

Expected project start date: 1st October 2013

Audio material is reproduced with a variety of devices, ranging from large multi-channel loudspeaker systems to portable devices and headphones. The widely used channel-based formats, used for audio production, storage and transmission, rely on the assumption that the reproduction system complies with the requirements of a standardized loudspeaker layout (stereo, 5.1, etc.) and provide one dedicated audio signal to each loudspeaker. Because of this intrinsic limitation of channel-based formats a new generation of audio formats, the so-called object-based formats, has been developed and its industrial application to audio production and delivery is rapidly growing.

Such formats rely on the description of the auditory scene by means of “audio objects”, each of which represents a separate component of the scene and is described by one or more audio files and additional metadata. In this context, a large number of audio objects may be effectively represented by a simple, point-like sound source. A different category of sound scene descriptors, which may still be regarded as objects, may not be effectively represented using the point-source model introduced above. This is the case, for example, of sound fields generated by a multitude of distributed sources (rain, traffic noise, etc.) or, as another example, of the reverberant field of a closed environment (concert hall, room, etc.). This category of objects is hereafter referred as “diffuse”.

Also Read  Eva Hoffmann and Frances Pearl Studentship Identifying novel cancer targets using computational techniques

The aim of this PhD project is to develop strategies to capture, represent and reproduce diffuse audio objects. The initial objective is to investigate the concept, definition and characterization of a diffuse object. The study will be carried out using objective descriptors of the sound field (pressure, intensity, correlation, etc.) that shall be analysed and interpreted in the light of the psychoacoustics of spatial sound perception. The project will then focus on the analysis of the future content delivery chain as currently developed by BBC R&D that may be extended for the diffuse object. More specifically, it is expected that new strategies will be developed for the capturing, data-efficient representation, and reproduction of the diffuse audio object.

The successful applicant will have a first-class honours MSc or equivalent degree in Engineering, Physics or Mathematics, with a strong background in signal processing, acoustics and audio technologies. A good knowledge of psychoacoustics and experience with sound engineering will be considered as an advantage.

The project is supported by the EPSRC’s Industrial CASE scheme.

Applicants must satisfy the relevant eligibility criteria, which due to funding restrictions include the condition that the applicant must have been an ‘ordinarily resident’ in the UK for 3 years prior to the start of the grant. http://www.epsrc.ac.uk/skills/students/help/Pages/eligibility.aspx

The interested candidates are encouraged to apply by sending their Curriculum Vitae and a cover letter to Dr Filippo Fazi ( [email protected] ).

To apply, please click the Apply button below

Leave a Reply

Your email address will not be published. Required fields are marked *