Statistics for Towards a general model for secure speech communications