r = NRMSE(a,b) computes the root mean squared error between two vectors normalised to the range of signal a. It is assumed signal a is the reference signal ARGUMENTS a ... 1st vector of data (reference signal) b ... 2nd vector of data RETURN r ... NRMSE between a and b NOTES - NaN removal assumes NaNs can only exist and start and end with continuous data in the middle e.g. NaN NaN NaN 1 2 54 65 765 34 NaN NaN NaN See also rmse