Combined speech coder and echo canceler

A method of processing a parametrically-encoded telecommunication signal for transmission from a near end station to a far end station, comprising steps of:

• receiving the parametrically-encoded signal at a send-input terminal;
• receiving a non-parametrically-encoded signal at a receive-input terminals;
• parametrically encoding the non-parametrically-encoded signal as a plurality of parametric components to provide an encoded receive-output signal at a receive-output terminal;
• providing at least one parametric component of the receive-output signal to a convolution processor;
• estimating an echo path impulse response between the receive-output and send-input terminals;
• synthesizing a component of the parametrically-encoded signal corresponding to said one component of the encoded receive-output signal;
• convolving said impulse response with said one component of the encoded receive-output signal to provide a first estimated echo signal;
• projecting the first estimated echo signal onto the parametric component of the synthesized signal to provide a second estimated echo signal;
• removing the second estimated echo signal from the synthesized signal to provide an error signal;
• and
• transmitting the error signal to the far end station.

An integrated parametric codec and echo canceler, comprising:

• a parametric encoder having an input terminal for receiving a first non-coded signal, a processing section for encoding the non-coded signal as a plurality of parametric signals, a synthesizer for producing a first synthesized signal in response to one of said parametric signals having a first parameter, and an output terminal for transmitting said parametric signals;
• a parametric decoder having an input terminal for receiving a parametrically-encoded signal having a second parameter, and configured for responsively producing a second synthesized signal;
• a convolution processor for generating an estimated echo signal on the basis of said first and second synthesized signals;
• conversion means for converting one of (i) said first synthesized signal in terms of said second parameter, and (ii) said estimated echo signal in terms of said second parameter;
• the conversion means connected with the convolution processor to provide said estimated echo signal in terms of said second parameter;
• and
• removal means for removing the estimated echo signal from the second synthesized signal.