Abstract:
A method of training a neural network to provide spatial parameters when stimulated by linguistic representation of speech, comprising providing thereto a linguistic representation of speech correlated with spatial parameters and a data storage medium having stored thereon a linguistic representation of speech correlated with spatial parameters for training such neural network.
Abstract:
The present invention teaches a method (400), device and system (300) utilizing at least one of: mapping a sequence of phones to a sequence of articulatory features and utilizing prominence and boundary information, in addition to a predetermined set of rules for type, phonetic context, syntactic and prosodic context for phones to provide a system that generates segment durations efficiently with a small training set.