Related Experiment Videos

Testing the assumptions of linear prediction analysis in normal vowels

M A Little1, P E McSharry, I M Moroz

  • 1Applied Dynamical Systems Research Group, Oxford Centre for Industrial and Applied Mathematics, and Pattern Analysis Research Group, Engineering Science, Oxford University, United Kingdom. littlem@maths.ox.ac.uk

Summary

Gaussian linear prediction analysis fails to capture all speech dynamics. Improved surrogate data testing reveals nonlinearities and non-Gaussianity are crucial for accurate speech modeling.

Related Concept Videos