Full duplex voice communication system and method

A full duplex voice communication method, the method comprising:

• receiving a near-end signal;
• receiving a far-end signal from a far-end communication device;
• estimating an acoustic echo within said received near-end signal responsive to said received far-end signal;
• cancelling said estimated acoustic echo from said received near-end signal;
• detecting whether, or not, a change has occurred in a near-end acoustic echo path;
• detecting whether, or not, said received near-end signal represents speech;
• and
• detecting whether, or not, said received far-end signal represents silence,
• wherein, responsive to a result of said near-end acoustic echo path change detection, a result of said near-end signal speech representation detection and a result of said far-end silence representation detection, the method further comprises:
• alternately attenuating frequency components of said echo cancelled near-end signal by a first frequency domain attenuation value and by a second frequency domain attenuation value, said second frequency domain attenuation value greater than said first frequency domain attenuation value;
• alternately attenuating a first function of said frequency component attenuated echo cancelled near-end signal by a first switchable attenuation value and by a second switchable attenuation value, said second switchable attenuation value greater than said first switchable attenuation value;
• and
• alternately attenuating a second function of said received far-end signal by a third switchable attenuation value and by a fourth switchable attenuation value, said fourth switchable attenuation value greater than said third switchable attenuation value.

A full duplex voice communication system comprising:

• a near-end input port arranged to receive a near-end signal;
• a far-end input port in communication with a far-end communication device and arranged to receive a far-end signal from said far-end communication device;
• an acoustic echo estimation functionality arranged, responsive to said received far-end signal, to estimate an acoustic echo within said received near-end signal;
• an acoustic echo cancellation functionality arranged to cancel said estimated acoustic echo from said received near-end signal;
• a frequency domain processing functionality;
• an echo path change detection functionality arranged to detect a change in a near-end acoustic echo path;
• a near-end speech detection functionality arranged to detect whether said received near-end signal represents speech;
• a far-end silence detection functionality arranged to detect whether said received far-end signal represents silence;
• and
• a system coordination unit,
• wherein said system coordination unit is arranged to:
• responsive to a lack of detection of said change in the near-end acoustic echo path by said echo path change detection functionality, detection that said received near-end signal represents speech and detection that said received far-end signal does not represent silence, control said frequency domain processing functionality to attenuate frequency components of said echo cancelled near-end signal by a first frequency domain attenuation value;
• and
• responsive to detection of said change in the near-end acoustic echo path by said echo path change detection functionality, detection that said received near-end signal represents speech and detection that said received far-end signal does not represent silence, control said frequency domain processing functionality to attenuate frequency components of said echo cancelled near-end signal by a second frequency domain attenuation value, said second frequency domain attenuation value greater than said first frequency domain attenuation value.