The synthesis of the aortic valve closure sound of the dog by the mean filter of forward and backward predictor

Abstract
The application of a new algorithm, the mean filter of forward and backward predictor, to synthesize the aortic component of the second heart sound (A2), according to an exponentially damped sinusoid model, is described. The resulting estimates of four modeling parameters composing each sinusoid of A2 (amplitude, damping factor, frequency, and phase) were used to synthesize the original aortic sounds. The synthesized sounds were then compared to the original sounds recorded on the thorax of six dogs and an error index was computed. The results show that the method is more precise than the forward predictor filter, the backward predictor filter, and the improved KT algorithm. The new algorithm is also highly stable and the error index between the original A2 and the synthesized waveforms is always less than 1%.