Browse Prior Art Database

Audio Response Unit

IP.com Disclosure Number: IPCOM000092377D
Original Publication Date: 1967-Dec-01
Included in the Prior Art Database: 2005-Mar-05
Document File: 2 page(s) / 23K

Publishing Venue

IBM

Related People

Nussbaumer, H: AUTHOR [+2]

Abstract

The audio response unit includes central processing unit CPU 1, digital to analog converter 2, low-pass filter 3, audio amplifier 4 and a loud speaker 5. Converter 2 is driven by groups of digital data. The latter is sent by CPU 1 either at the pitch frequency for the voiced sounds or at random rate for unvoiced sounds. Converter 2 converts the binary data into PAM pulses, the amplitudes of which depend on the sound to be reproduced through amplifier 4 and loudspeaker 5.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 71% of the total text.

Page 1 of 2

Audio Response Unit

The audio response unit includes central processing unit CPU 1, digital to analog converter 2, low-pass filter 3, audio amplifier 4 and a loud speaker 5. Converter 2 is driven by groups of digital data. The latter is sent by CPU 1 either at the pitch frequency for the voiced sounds or at random rate for unvoiced sounds. Converter 2 converts the binary data into PAM pulses, the amplitudes of which depend on the sound to be reproduced through amplifier 4 and loudspeaker 5.

A voice analysis gives information necessary for voice synthesis. This information is the aggregate function and the excitation function. The data corresponding to the two functions are in the memory of CPU 1. The aggregate function gives information about the amplitude of each sample of speech. The excitation function gives information about the decision of pitch no-pitch necessary to know if a sound is voiced or unvoiced and the value of the pitch if the sound is voiced.

The audio response unit uses, to synthesize the sounds, a symmetrical data frame corresponding to a tapped filter. CPU 1 simulates the response of the filter by using a program. For example, if the tapped filter has 30 taps with an 8 bits per data definition, for every sample of speech the audio response unit needs 8 x 30 = 240 bits or 30 bytes. Consequently, the output of CPU 1 to converter 2 is 30 bytes for each sample of speech.

These bytes carry the aggregate function corresponding to the spectrum functi...