Digital Equipment Corporation

Brian Scott Eberman

Pedro J Moreno

Oren David Glickman

In a computerized method for processing speech signals, first vectors representing clean speech signals are stored in a vector codebook. Second vectors are determined from dirty speech signals. Noise and distortion parameters are estimated from the second vectors. Third vectors are predicated, based on estimated noise and distortion parameters. The third vectors are used to correct the first vectors. The third vectors can then be applied to the second vectors to produce corrected vectors. The corrected vectors and the first vectors can be compared to identify first vectors which resemble the corrected vectors.

Environmently compensated speech processing

In a computer system, a user interacts with a plurality of applications using natural language text provided by an input device. In order to enable the applications for user interaction, rules are registered on a blackboard of an arbiter. The arbiter can be localized, or distributed. Each rule is identified with a particular application, and can include a callback procedure. Each rule includes a left-hand side symbol and a right-hand side string which is a translation of the symbol. A parser parses a natural language input text string using the register rules to generate an evaluation tree. The evaluation tree is expressed in the form of an interpretive language. The evaluation tree is interpreted to produce at least one application request. The application request is generated by the callback procedure associated with a specific one of the rules applied during the parsing of the input text. The request is transmitted to the application identified with the applied rule.

Growing community of inventors

Somerville, MA, United States of America

Brian Scott Eberman