Automatic Speech Segmentation
This worksheet demonstrates the use of the Forward-Backward Divergence model (FBD) in Automatic Speech Segmentation, and how it detects discontinuities in the voice signal. It illustrates in the example below how it is possible to enlarge some segments of the speech (vowels enlargement for instance). To realize this result, it is possible to visually and acoustically perceive the stationary segments of the speech signal.