Some of the most recent hearing aids are featuring acoustical beamforming able to artificially narrow the directivity of the external microphone so that only the signals in front of the wearer are amplified in certain situations. This could be of great use if the wearer is in a one-to-one conversation in a noisy environment since it greatly improve the signal-to-noise ratio of the remote speaker, who is assumed to be facing the wearer of the hearing aid. Such an acoustical beamformer is achieved by using a microphone array of -typically- two microphones close to each other on the outside of the hearing aid device.
The researchers for the Sonomax-ÉTS Industrial Research Chair have been working on the implementation of such beamforming approach on a pair of digital hearing protectors able to let speech and warning signals get through. In such case, only one microphone is available on the outside of each digital hearing protection device, but since left and right protectors (binaural) are electrically wired and it is therefore possible to reasonably increase the signal-to-noise ratio of a remote speaker.
The proposed research project deals with the refining of the signal processing algorithms to further improve the signal-to-noise ratio of the beamformer and to reduce the front/back confusion created by the side lobes of the proposed binaural setup. In particular, the research project should evaluate if the protected acoustical signal picked-up by the internal microphones (each digital protector has one external and one internal microphone) could be utilized for a sub-band adaptative beamforming approach.
The selected student should be fully knowledgeable in acoustical signal processing and electro-acoustical systems. The selected student should also have a good experience in Matlab or similar signal processing software and should be somehow familiar with DSP implementation. Prior experience in microphone array would be a plus.
The selected student would be responsible for the development, in Matlab, of a signal processing algorithm able to achieve a directional beamformer using binaural microphones and for the refinement of this method by including the two extra internal microphones for reduced side-lobes. The selected student would then be provided tools and training for DSP implementation and would be responsible for the implementation the proposed approach. The evaluation of the proposed approach would be conducted on Artificial Test Fixtures in collaboration with the graduate student.
The selected student would be interacting on a daily basis with the chair holder and one of the graduate student of the industrial research chair; he/she may also be in contact with the Sonomax engineering team for very specific aspect of the hardware design of the external microphones.