We currently have three final-year projects running on some parts in a hearing-aid devices or perhaps more generally in an apparatus that deals with audio and speech processing.
Within the current projects, the students study a phoneme detector, an auditory scene analysis algorithm (mainly to identify and characterize noise in audio) , and a sigma-delta D/A converter to produce the noise at the boundary of the FPGA.
We are looking for more students to look at other parts, such as
- an input sigma-delta ADC to sample the sound/speech
- Matlab/C++/Python software to run a hardware-in-the-loop system
- a digital filterbank/wavelet for time-frequency analysis of the signals
- an ultra-low power analog SC filterbank/wavelet for time-frequency analysis of the signals
- an ultra-low power bandgap reference
- an ultra-low power input audio buffer and anti-aliasing filter
The ambition is to build an FPGA demonstrator and eventually also an ASIC.