Browse Prior Art Database

Vocoder Corrections Disclosure Number: IPCOM000076365D
Original Publication Date: 1972-Feb-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 2 page(s) / 13K

Publishing Venue


Related People

Nemeth, A: AUTHOR [+1]


Described is an improvement in vocoder correction methods, by use of versatile protheses sets.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 2

Vocoder Corrections

Described is an improvement in vocoder correction methods, by use of versatile protheses sets.

In vocoders, i.e. devices storing signals which represent, in coded form words destined to be reproduced in an audible way, each sound is represented by two sequences of signals, which define respectively for given time intervals: - the energy value in a number of frequency bands (excitation function) - the value of the "pitch" (in case of voiced sounds) or the absence of pitch (in case of unvoiced sounds).

These values are provided by a speech analyzer. However, some sounds cannot be accurately reproduced, due in particular to the lack of sharpness of the analysis, from which it results that some components are missed or chopped. In order to correct these defects, it is necessary to replace the representation of some sounds provided by the analyzer, by artificially built representations (operation called "prosthesis").

In spite of the use of prostheses, the quality of the sounds remains poor. The purpose of the system described hereafter is to improve this quality.

The system is based on the results of studies showing that the structure of a sound is strongly influenced by its environment, and that, consequently, a prosthesis suitable in some words for replacing a given sound, gives bad results when it is used, in other words, for replacing the same sound. For example, in the English language, the unvoiced plosives are aspirated in stressed portions (Ten) but not in unstressed positions (leTTer) and the prosthesis of the letter T must be fitted to the following voiced sound. Another example relates to some voiced sounds, especially the nasal sounds and the voiced fricatives. These sounds are no...