An audio system is described including an audio processor, an amplifier and a speaker and a distortion estimator. The distortion estimator calculates at least one of an expected response of the amplifier to an audio signal and an expected response of the loudspeaker to an audio signal. The distortion estimator is operable to generate a distortion prediction signal determined by a difference between an expected non-linear response of at least one of the loudspeaker and the amplifier and an expected linear response of at least one of the loudspeaker and the amplifier. A controller coupled to the audio processor and a control input of the audio processor may vary the operating parameters of the audio processor depending on the estimated distortion.