INVESTIGATION AND EVALUATION OF THE GEL SPEECH SYSTEM TEST SET

Abstract

Mathematical models of the articulation scores of an untrained and a fully trained team of listeners were developed. The maximum variance of the trained team was found, and the precision of the results was computed. An investigation was made of the GEL Speech System Test Set's capability for providing a measure of the degradation in intelligibility that occurs when a voice signal is passed through a noisy speechcommunication channel. The test set's measure of intelligibility was compared with that of a trained articulation team, and calibration curves were established for a variety of interference conditions. The validity of the test set design principle was verified for particular interference types, and the ability of the test set to perform in accordance with that design principle was determined. The inherent test set limitations are briefly discussed, and the possible extension of the test set's capability for ranking communication systems for other interference conditions is summarized. (Author)

Document Details

Document Type
Technical Report
Publication Date
Feb 28, 1961
Accession Number
AD0256258

Entities

People

  • C.w. Stuckey
  • D.w. Robertson

Organizations

  • Georgia Tech

Tags

DTIC Thesaurus Topics

  • Calibration
  • Communication Systems
  • Degradation
  • Intelligibility
  • Mathematical Models
  • Models
  • Precision
  • Speech
  • Test And Evaluation
  • Test Sets

Fields of Study

  • Engineering

Readers

  • Aerospace Test and Evaluation
  • Regression Analysis.
  • Speech Processing/Speech Recognition.