The results shown in the Figure 9 need to be carefully interpreted. Although they show that the performance of the MG algorithm is worse for large α, it is also true that such large values are not used in practice. The purpose of α is to select the amount of noise to be removed, so there is no point in choosing either a very large or very small α. In Table 4 we illustrate the effect that different values of α have on the PSNR and the number of FAS cycles used by our algorithm to solve the problem. The maximum PSNR was for small α = 0.005, and only 8 FAS cycles were needed.