Acoustic echo cancellation can be used to remove talker feedback in hands-free systems. Fast convergence and good tracking capabilities cannot be achieved by classical transform domain adaptive filtering algorithms when the reference signal has a variable rank autocorrelation matrix. During the low rank phases of the speech signal, some of the transform-domain tap coeficients become irrelevant to the adaptation process and stop adapting. When the autocorrelation matrix gains full rank, there will be no longer any \frozen" weights. In this paper, we focus on the DCTLMS algorithm and present a new method using a DCT based delay estimate from other coeficients to move the frozen weights closer to the optimal point and, consequently, reduce the overall re-convergence time.