720834.fig.0016
Figure 16: Comparison of average precision for different voice datasets and different preprocessing methods.