Browse Prior Art Database

Speech Recognition System Using Formant Transient Detection

IP.com Disclosure Number: IPCOM000096860D
Original Publication Date: 1963-Dec-01
Included in the Prior Art Database: 2005-Mar-07
Document File: 2 page(s) / 48K

Publishing Venue

IBM

Related People

Clapper, GL: AUTHOR

Abstract

The arrangement detects transients in a speech recognition system where such transients are utilized in subsequent analysis of the speech sounds. Square wave pulses representative of local maxima in the speech energy spectrum are supplied by the input lines M1...M9 to a plurality of differentiators such as DF1...DF9 to create a pulse at the end of the local maxima.

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 100% of the total text.

Page 1 of 2

Speech Recognition System Using Formant Transient Detection

The arrangement detects transients in a speech recognition system where such transients are utilized in subsequent analysis of the speech sounds. Square wave pulses representative of local maxima in the speech energy spectrum are supplied by the input lines M1...M9 to a plurality of differentiators such as DF1...DF9 to create a pulse at the end of the local maxima.

These differentiated pulses combine with neighboring active maxima to set latches such as L2, L3, L5, L6, etc., to represent transient conditions designated by an R for a rising transient and an F for a falling transient. if a local maxima ends and neither of its associated transient latches is set, e.g., either L2 or L6 or both, for input M2, a third latch indicating the steady condition is set, e.g., L4.

The Nor's used in the inputs to the steady condition latches make it possible to have all latches similar. The second differentiators D2F1, D2F2, etc., operate to delay setting of the S latches, such as L1 and L4, until after the transient latches have had time to set. Output lines indicating the transient conditions detected in local maxima such as M2R, M2S and M2F are connected to suitable circuits for analyzing various combinations of signals to provide recognition of the input speech signals.

1

Page 2 of 2

2

[This page contains 3 pictures or other non-text objects]