Browse Prior Art Database

Speech Data Adaptive Multiplexer

IP.com Disclosure Number: IPCOM000043043D
Original Publication Date: 1984-Jul-01
Included in the Prior Art Database: 2005-Feb-04
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Callens, P: AUTHOR [+3]

Abstract

In a system wherein digitized speech and pure data are to be multiplexed and transmitted by blocks (e.g., HDLC or SDLC) of bits, one has to provide for means for optimizing each block's contents while assigning the highest priority to speech-related bits. Most such systems are based on voice activity detection (VAD) information. In other words VAD information is used to control the switching from one block format, wherein most bit positions are assigned to speech whenever speech is detected by a VAD device, to another block format assigning all bit positions to data whenever speech is absent. One bit per block is used under these circumstances to characterize each block. The system can still be optimized by taking into account specific characteristics of the speech signal.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 80% of the total text.

Page 1 of 1

Speech Data Adaptive Multiplexer

In a system wherein digitized speech and pure data are to be multiplexed and transmitted by blocks (e.g., HDLC or SDLC) of bits, one has to provide for means for optimizing each block's contents while assigning the highest priority to speech-related bits. Most such systems are based on voice activity detection (VAD) information. In other words VAD information is used to control the switching from one block format, wherein most bit positions are assigned to speech whenever speech is detected by a VAD device, to another block format assigning all bit positions to data whenever speech is absent. One bit per block is used under these circumstances to characterize each block. The system can still be optimized by taking into account specific characteristics of the speech signal. For instance one can add to the VAD information, so-called voiced/unvoiced characteristic information, to again switch from one block format to another. A block format for a voiced speech signal will be made to contain a bit frame devoted to speech, larger than the bit frame devoted to unvoiced speech. Several means have already been implemented for discriminating between voiced and unvoiced speech signals. Finally, a further improvement on block formatting can be achieved when using a split-band Voice Excited Predictive Coder (VEPC) for performing the A/D (analog-to-digital) speech conversion. With this type of coder, the residual base-band signal is split into...