Abstract: An analog adaptive filter for deconvolving distorted signals. A clipped vion of the least mean square algorithm is implemented in a first embodiment by placing an acoustic surface wave storage correlator in a feedback loop and performing an iterative two step process to adjust the weight function of the correlator to make the output of the filter converge to a desired signal. In the first step, the distorted signal is applied to the top plate port of the storage correlator and the filter output from the acoustic port is delayed and subtracted from a training signal to produce an error signal. In the second step, the error signal is applied to the top plate port and the distorted signal is applied to the acoustic port which results in the correlation of the acoustic signal and the plate signal being added to the weight function stored in the storage correlator. The two step process is repeated until the filter output from the plate port converges to an approximation of the training signal.