720834.fig.0017
Figure 17: Comparison of average recall for different voice datasets and different preprocessing methods.