In certain systems it is very important to calculate quickly an inverse filter that has a large number of taps. Several approaches to solving normal equations are investigated. One fast method uses conjugate gradients in the frequency domain. To enhance this method a fast method that calculates only linear convolutions in conjugate gradients by FFT is proposed. Simulation results show that the sum of the squares of the residual errors of our method is as low as that of the Levinson algorithm, which is slower than the method proposed.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.