Browse Prior Art Database

Data Reduction Method for Voice Excited Vocoder

IP.com Disclosure Number: IPCOM000090767D
Original Publication Date: 1969-Jul-01
Included in the Prior Art Database: 2005-Mar-05
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Levilion, ME: AUTHOR [+2]

Abstract

In this coding method, the amount of information required by a voice-excited vocoder VEV synthesizer can be reduced by a factor of about 13% on the average. A description of the standard coding method is first given.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 75% of the total text.

Page 1 of 1

Data Reduction Method for Voice Excited Vocoder

In this coding method, the amount of information required by a voice-excited vocoder VEV synthesizer can be reduced by a factor of about 13% on the average. A description of the standard coding method is first given.

Two basic series of information must be coded for a VEV. The first series corresponds to the baseband signal. The second series corresponds to the upper energy channels. In one high fidelity VEV, the baseband requires sampling at 1400 Hz, with 6 bits of information per sample. Each channel requires sampling at about 50 Hz with 4 bits per sample. As a result, the following coding system is used. One channel sample is associated with two baseband samples. This results in a group of 16 bits which can be distributed in a set of two 8-bit bytes suitable for transmission on a channel. In a coded voice signal, the channel energy, as measured by the VEV analyzer, is often equal to zero, resulting from the fact that the energy of the voice is concentrated in the formant areas of the spectrum.

This coding method reduces the amount of information required whenever two adjacent channels show zero energy level, while still preserving a matching with the 8-bit bytes frame.

A group of four baseband samplings is transmitted first corresponding exactly to 3 bytes. One of these bits is reserved for signalling purposes, thus slightly reducing the coding quality on one of the baseband samplings. This bit is used to indicat...