The Measurement of Speech Intelligibility.

Abstract

A model of the speech recognition process was presented in an attempt to find adequate measures of intelligibility and intelligence. An experiment was designed to test the model and the validity of the measures. Four trained subjects responded to distorted speech samples from a message set containing twelve single words and from a message set containing thirty-six three-word sequences. The goal was to predict the scores which would be obtained on the sequences from the scores obtained on the single words. A posteriori predictions were attempted using percentage correct measures and measures derived from information theory and a priori predictions were made from the decision theory model proposed in the study. The results of these data support the information theory and decision theory model. (Author)

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1970
Accession Number
AD0720091

Entities

People

  • Wilson P. Tanner

Organizations

  • University of Michigan

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Automated Speech Recognition
  • Decision Theory
  • Human Factors Engineering
  • Information Theory
  • Intelligibility
  • Measurement
  • Recognition
  • Sequences
  • Speech

Fields of Study

  • Psychology

Readers

  • Acoustics.
  • Brain and Cognitive Science; Experimental Psychology; Cognitive Neuroscience
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Translation