SCRUDRIVA – Sound field control using head-worn multi-driver arrangements

• Applicant: Professor Dr. Jürgen Peissig
• Subject Area: Acoustics
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 562226092

The research project SCRUDRIVA (Sound field control using head-worn multi-driver arrangements) deals with the possibilities, challenges and limitations of recreating a natural sound field within a hemispherical head-worn multi-driver arrangement in order to improve the reproduction of spatial sound scenes. Therefore, a setup is used which, unlike conventional stereo headphones, is equipped with up to 32 loudspeakers. With the advance of virtual and augmented reality applications, the reproduction of spatially perceived audio stimuli via headphones is becoming increasingly important. This is currently predominantly realized with binaural rendering, in which the stimulus is convolved with a head-related transfer function (HRTF) for the direction of sound incidence to be virtually represented and then reproduced via frequency response equalized headphones. The best results can be achieved with this approach if individual HRTFs are used that have been measured in the ear canal close to the eardrum. However, the metrological determination of these is complex, which is why generic HRTFs, e.g. from artificial heads, are often used. In addition, there are many approaches for applying individual cues to a generic HRTF in order to obtain a semi-personalized HRTF. In any case, the use of non-individual HRTFs leads to a less realistic perception of the sound scene, which often manifests itself in problems such as front- back confusion and lack of externalization. Our research project is addressing a new method of binaural synthesis that is not based on measured or simulated HRTFs or binaural room impulse responses (BRIRs), but reconstructs the natural sound field generated by a far-field source around the pinna by linear superposition of the sound fields generated by each loudspeaker of a multi-driver arrangement. It is hypothesized that this reconstruction subsequently leads to a natural interaction between the sound field and the pinna and thus to an inherent personalization of the listening experience. To investigate this hypothesis, simulation-based signal processing strategies for sound field reconstruction will be developed using a digital virtual twin, which will then be evaluated both metrologically with a sound intensity probe and perceptually in listening tests. It is postulated that the use of a multi-driver arrangement and the focus on the reconstruction of the three-dimensional sound field may result in notable enhancements in comparison to binaural rendering with non-individual HRTFs.

• DFG Programme: Research Grants
• Co-Investigator: Dr.-Ing. Stephan Preihs