AES E-Library

AES E-Library

A Comparison of Two Types of Digitized Autocorrelation Vocoders

Document Thumbnail

Intelligibility test results are presented for two types of 2400 bit/second autocorrelation vocoders. One vocoder directly transmits the digitized coefficients in a Laguerre Expansion of the autocorrelation function. The other vocoder contains a resistor matrix which transforms these coefficients to spectrum amplitude values and is therefore compatible with present channel vocoders. A comparison of the effects of fixed versus variable reference spectrum encoding as well as a taped demonstration will be presented. This paper describes the results of some tests of the intelligibility of two forms of autocorrelation vocoders digitized at 2400 bits/second. One of these vocoders (V-1) is called a Formant Synchronous Autocorrelation Vocoder and was described in Reference 1. The analyzer of this vocoder is shown in Figure 1. The configuration is such that the short time autocorrelation function of the speech signal is expanded in terms of a set of sixteen Laguerre functions. A set of bandpass filters, covering the three formant frequency ranges, each followed by a limiter is used to normalize the energy in the three formant bands so as to reduce the effects of spectrum squaring normally inherent in autocorrelation analysis. Each of these amplitude-limited formant signals forms the reference input for a set of 16 synchronous detectors, one at each tap on the Laguerre filter.

Authors:
Affiliation:
AES Convention: Paper Number:
Publication Date:
Permalink: https://www.aes.org/e-lib/browse.cfm?elib=993

Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!

This paper costs $33 for non-members and is free for AES members and E-Library subscribers.

Learn more about the AES E-Library

E-Library Location:

Start a discussion about this paper!


AES - Audio Engineering Society