Aspects of Speech Recognition by Computer
Department of Computer Science, Stanford University., 1969 - Automatic speech recognition - 210 pages
The paper describes techniques and methodology which are useful in achieving close to real-time recognition of speech by a computer. To analyze connected speech utterances, any speech recognition system must perform the following processes: preprocessing, segmentation, segment classification, recognition of words, recognition of sentences. The paper presents implemented solutions to each of these problems which achieved accurate recognition in all the trial cases. (Author).
What people are saying - Write a review
We haven't found any reviews in the usual places.
Preprocessing for Speech Analysis
7 other sections not shown
Other editions - View all
A-D converter acoustic adjacent segments algorithm amplitude analog-to-digital converter average parameters candidate list candidate selection process candidate space capacitor Chapter circuit command connected speech utterances corresponding decoding defined described detected digitized display duration eh7 OK OK error recovery feature matrix Figure flowchart Formant frequency fricative segments function grammar HAND-EYE-EAR system hardware preprocessor heuristic Hz db incoming message input kHAT labeled language lexicon M3DIUW BLOCK machine language mapping microphone minimal segments multiplexer NEHi OCTAL OK EH OK OK OK operational amplifier output PASS performed phoneme phoneme groups PICK UP BIG PICK UP SMVLL primary segmentation print-name problem R3SCVN Read pulse real-time REAO recognition system reset results obtained RIGHT CORNER routines S3GW3NT secondary segmentation segmentation procedure similarity evaluation procedure similarity score SMVLL BLOCK sound speakers speech recognition stored subroutines sustained segment SWVLL syllable terminal symbols TH3N TOP L3FT CORN3R transitional segments vocabulary vowel word recognizer zero-crossing