3D Audio
Our research activity concerns the development of innovative techniques for spatial audio synthesis with particular attention to binaural audio synthesis and real-time audio rendering. We analyze the contribution of the external ear (pinna, PRTF) in relation to personal HRTF, proper to each individual, and model the physical features that have a perceptual interest for vertical localization of sound. The two main objectives are:
- the development of efficient real-time algorithms for spatial sound rendering, for a coherent simulation of complex multi-source acoustic environments, where both the listener and the sound source are expected to move in dynamic conditions;
- the design of customized models with anthropometric measures as input parameters.
Customized HRTF models
The long-term goal is the implementation of structural models of head-related transfer functions (HRTFs) for binaural synthesis that approximate the real behaviour of the auditory space, to be applied in a real-time context to any subject.
We have designed an algorithm for the decomposition of pinna-related transfer functions (PRTFs) into ear resonances and frequency notches due to reflections over pinna cavities and following that we set up a method for extracting the frequencies of the most important spectral notches. Ray-tracing on different pinna images reveals a convincing correspondence between extracted frequencies and pinna cavities. The proposed model for PRTF synthesis allows to control the behaviour of resonances and spectral notches separately through the design of two distinct filter blocks, and is suitable for integration into a structural HRTF model and for parametrization over anthropometrical measurements.
Mixed structural modeling
We propose a novel framework for synthetic HRTF design and customization, that combines the structural modeling paradigm with other HRTF selection techniques: namely, the Mixed Structural Modeling (MSM) approach regards the global HRTF as a combination of structural components, which can be chosen to be either synthetic or recorded components. In both cases, customization is based on individual anthropometric data, which are used to either fit the model parameters or to select a recorded component within a set of available responses.
The research process aims at building a completely customizable structural model through subsequent refinements, starting from a selection of recorded HRIRs to a totally synthetic filter model. The intermediate steps are balanced mixtures of selected pHRIRs and synthetic structural components.
Application areas
A bunch of example application areas of 3D audio technologies are:
- Rehabilitation and education of blind or visually impaired users; such context is particularly significant where the use of non-visual modalities stimulate interaction with a safe environment. The haptic and auditory modalities can be exploited to design interfaces for non-sighted users, e.g. to render spatial information in non-visual exploration of maps.
- The design of a wireless headphone device equipped with sensors able to detect the relevant parameters needed to fine tune the HRTF model both before and during listening (motion sensors, multiflash camera).
Research Threads
3D Audio
Audio in multimodal interfaces
Audio restoration
Interactive environments for learning
Music expression modeling
Physically-based sound modeling
Virtual rehabilitation
History of CSC research
Sub-Threads
Customized HRTF models
Mixed structural modeling
Application areas
Projects
A complete list of projects and industrial partners can be found here.