Device-optimized perceptual enhancement of received speech for mobile VoIP and telephony

A solution to the remote speech enhancement in presence of local background noise is fundamental in mobile communications. The toolset presented in this paper proposes a new way of adapting the signal received from the remote peer in accordance to the surrounding noise. Based on a real-time analysis of the background noise, it selectively boosts critical frequency bands to enhance the intelligibility of the received signal. Automatic gain adjustment and particular care on mobile devices computational constraints make the solution effective in terms of power and resource consumption.