Two cooperative methods for the segmentation of running speech