Right now we compute the mean and standard deviation for weighted and unweighted metrics individually, however they use the same stats. We can improve the efficiency by only computing mean and standard deviation once for each pair of weighted and unweighted metrics, i.e. MFCCs and Weighted MFCCs.