An audio apparatus including a circuit for applying a dynamic range adjustment processing to an input signal to generate an output signal for rendering by a loudspeaker. The dynamic range adjustment processing is dependent on a set of dynamic range adjustment parameters. A first linear echo-cancellation filter generates a first compensated signal from the input signal and a first adapter determines a set of filter parameters for the first linear echo-cancellation filter in response to the first compensated signal and a microphone signal. An audio circuit generates a second compensated signal by applying the dynamic range adjustment processing and a second echo-cancellation filter to the input signal where the second echo-cancellation filter corresponding to the first echo-cancellation filter. A second adapter determines the set of dynamic range adjustment parameters in response to the second compensated signal and the microphone signal.