公开(公告)号：US3557310A

公开(公告)日：1971-01-19

申请号：US3557310D

申请日：1968-07-03

Applicant: FARRINGTON MFG CO

Inventor： PRESTI ANTHONY J ,  KERSTA LAWRENCE G

IPC: G10L15/00 ,  G10L1/08

CPC classification number: G10L15/00

Abstract: Speech recognition equipment is disclosed which distinguishes the sounds ''''OH'''' and ''''ONE'''' (respectively corresponding to the numerals 0 and 1). The recognition equipment takes advantage of the fact that the characteristic frequency of the second formant of ''''OH'''' increases with the time while the characteristic frequency of the corresponding formant of ONE decreases with respect to time. The sound to be recognized is fed to a frequency discriminator after a test has been made to distinguish the sound from background noise. The discriminator output is applied to two sampling circuits, one of which samples the discriminator output signal shortly after the presence of the sound is detected and the second of which samples the signal shortly after the first sample is taken. The samples are held in capacitive storage means. After both samples have been taken, the outputs of the capacitors are applied to a subtractor. Thus, the polarity output of the signal from the subtractor indicates whether the sound was ''''OH'''' or ''''ONE.