Develop a phone or an edge application for active noise cancelling, where the source of noise is a person and represented by an audio recording. The implementation should comprise a method for the recording of the noise source, feature-based analysis of the noise spectrum (FFT, etc.), signal generation and signal rendering to headphones. Ideally, the implementation platform should be an Android phone. If not otherwise possible, it might be a (GPU enhanced) embedded hardware. If performed as a diploma thesis, must include user-based evaluation. Requires good skills in signal processing and (embedded) programming.
