view Code/Descriptors/yin/private/src/minparabolic.m @ 2:5fd388fdd6ef

initial commit - this file allows the programmer to select which of the PRAAT generated metrics the user wishes to use for classification, and then applies the k-means classifier.

% [y,idx] = minparabolic(x) - locate minima using parabolic interpolation
%
%  x: input vector
%  y: same as x, with local minima replaced by interpolated minima
%  idx: position of interpolated minima (fractionary)
%
% At each local minimum of x, that sample and its neighbors are used to
% determine a parabola.  The minimum of the parabola is used as the new
% minimum value, the abscissa of the minimum is used as its position.