Browse Prior Art Database

Vowel Separation by Time Ratio Measurements

IP.com Disclosure Number: IPCOM000096263D
Original Publication Date: 1963-Mar-01
Included in the Prior Art Database: 2005-Mar-07
Document File: 2 page(s) / 24K

Publishing Venue

IBM

Related People

Harper, TP: AUTHOR

Abstract

This system separates components of signals, such as voicing waveforms, by asymmetry detection. The system can also separate frictional from voiced sounds. This is due to the fact that frictional sounds are symmetrical in nature and register nothing on this asymmetry detector. The method of detecting the occurrence of vowel sounds is by generating positive or negative pulses. These are representative of the time between axis crossings. This polarity is dependent upon whether the crossing is positive-going or negative-going. These pulses are integrated and added algebraically to derive a summing output signal. This is representative of a time -ratio, i. e., period of positive wave over period of negative wave.

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 78% of the total text.

Page 1 of 2

Vowel Separation by Time Ratio Measurements

This system separates components of signals, such as voicing waveforms, by asymmetry detection. The system can also separate frictional from voiced sounds. This is due to the fact that frictional sounds are symmetrical in nature and register nothing on this asymmetry detector. The method of detecting the occurrence of vowel sounds is by generating positive or negative pulses. These are representative of the time between axis crossings. This polarity is dependent upon whether the crossing is positive-going or negative-going. These pulses are integrated and added algebraically to derive a summing output signal. This is representative of a time -ratio, i. e., period of positive wave over period of negative wave.

According to this time-ratio, the presence of wave asymmetry is detected and, through this, the occurrence of Voicing, as well as particular kinds of Voicing having characteristic periods and time ratios.

The system has a Schmitt trigger for pulse generation, a pair of opposing diodes with integrators in parallel for polarity separation, and a slide wire potentiometer for ratioing. The input is a microphone M and amplifier A. These present a transduced voice wave to Schmitt trigger ST. This generates a different pulse each time the speech wave crosses the reference axis. The polarity of the pulse depends upon whether the crossing is negative-going or positive-going. Resulting positive pulses are detected by diode D...