Evolved frequency log-energy coefficients for voice activity detection in hearing aids
Roberto Gil‐Pita, Joaquín García-Gómez, Marta Bautista-Durán, Elías F. Combarro, Alberto Cocaña-Fernández · 2017
Eco-efficiency in hearing aids is an important issue, related to the maximization of the battery life. In order to minimize the power consumption, the embedded digital signal processor works at very low clock rates, constraining the implementation of signal processing techniques. The implemented algorithms can only use a small number of instructions per second and a small amount of memory. One of the main algorithms implemented in nowadays hearing aids is the voice activity detection algorithm, useful for several noise reduction and speech enhancement algorithms. The objective of this paper is the study of the implementation of voice activity detection algorithms in hearing aids using tailored fuzzy features, taking into account the optimization of the available resources.