Speech
Abstract
Programs for processing speech waveforms to yield spectral analysis and phonetic segmentation have been under development on the FDP-Univac 1219 facility. Available outputs now include: (1) homomorphic (cepstrally smoothed) spectra; (2) a set of parameters extracted from the spectra; (3) some preliminary segmentation markers; (4) linear predictive coding coefficients, spectra, and formant estimates; and (5) a fundamental frequency estimate. Current efforts are directed toward refining the segmentation algorithm. Implementation of the supporting software for the TX-2 Speech Data Base is proceeding. The TSP(Terminal Support Processor) hardware is now operating well enough to pass acceptance test, and emphasis is shifting from hardware checkout to system programming.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 31, 1972
- Accession Number
- AD0745970
Entities
People
- James W. Forgie
Organizations
- Massachusetts Institute of Technology