/ Features / Fundamentals / Production / Tutorials / Software / Home
Overview:
Section 3: Overview

Speech recognition begins when a person speaks into an electronic device, such as a microphone. The microphone converts the speech to an analog signal, which is then converted to a digital signal to be stored on a computer. A speech recognizer, a program running on the computer, attempts to decode the meaning of the speech signal. Certain features of a person's speech are needed by the speech recognizer to decode the signal. These features must be numerically measured and stored in a form the recognizer can understand. The process of measuring the features needed by a speech recognizer is called feature extraction. Continue to Section 3.1 for a theoretical overview of the feature extraction process.

Contents:
Section 3: Feature Extraction

3.1   Overview
  3.1.1   What is a Feature?
  3.1.2   Frame-Based Processing
  3.1.3   Simple Transformations

3.2   Signal Flow Graphs
  3.2.1   Time Domain Analysis
  3.2.2   Frequency Domain Analysis
  3.2.3   Recipes
  3.2.4   More Complex Recipes

3.3   Rapid Prototyping
  3.3.1   Getting Started
  3.3.2   Transform Builder Interface
  3.3.3   Absolute MFCC Computations
  3.3.4   Algorithm Components

3.4   Isip_Transform
  3.4.1   Command Line Interface
  3.4.2   Command Line Options

3.5   MFCC Example
  3.5.1   Basic MFCCs
  3.5.2   Enhancements

3.6   Components
  3.6.1   Infrastructure
  3.6.2   Statistics and General Mathematics
  3.6.3   Signal Processing

3.7   Command Synopsis
Section 3: Feature Extraction
   
Table of Contents   Section Contents   Previous Page Up Next Page
      Glossary / Help / Support / Site Map / Contact Us / ISIP Home