To evaluate the performance of the proposed tuning method in non-stationary noise conditions, we conduct an experiment with the same setup as in Fig. 2. It can be seen from Fig. 7 (a) that the adjusted Ko(n) by the proposed method shows significantly less bias and closely similar to that of the oracle Ko(n). Specifically, it main- Figure 7: Comparing Ko(n) trajectories corresponding to the AKF- Oracle method, proposed method , AKF-RMBT method [19], and AKF-SMBT method [20], where the utterance sp27 is corrupted with 5 dB: (a) babble and (b) factory noises.