Patent
1992-11-30
1994-11-01
Knepper, David D.
395 218, 395 243, G10L 302, G10L 904
Patent
active
053613244
ABSTRACT:
An analyzer performs spectral analysis of a characteristic parameter of an input speech signal per unit time so as to calculate, as a recognition parameter, orthogonal expansion of a spectrum of the analyzed characteristic parameter. A formant frequency of the input speech signal is detected, and a compensation value is calculated for the recognition parameter using the formant frequency. The recognition parameter is compensated in accordance with variations of the input speech signal caused by a noisy environment using the compensation value. The compensated recognition parameter is compared with reference parameters to output a recognition result.
REFERENCES:
patent: 4461024 (1984-07-01), Rengger et al.
patent: 4918735 (1990-04-01), Morito et al.
patent: 4922539 (1990-05-01), Rajasekaran et al.
patent: 4933973 (1990-06-01), Porter
patent: 4945568 (1990-07-01), Willems
patent: 4956865 (1990-09-01), Lennig et al.
patent: 5001761 (1991-03-01), Hattori
patent: 5040217 (1991-08-01), Brandenburg et al.
Furui, "Cepstral Analysis Technique for Automatic Speaker Verification," IEEE Trans. on ASSP, vol. ASSP-29, No. 2, Apr. 1981, pp. 254-272.
ICASSP'87 (1987 International Conference on Acoustics, Speech and Signal Processing, Dallas, Apr. 6th-9th 1987), vol. 2, pp. 717-720, IEEE New York, U.S.; Y. Chen "Cepstral domain stress compensation for robust speech recognition"; pp. 718, 719: Cepstral domain stress compensation-A hypothesis driven approach.
Hamada Masahiro
Takizawa Yumi
Knepper David D.
Matsushita Electric - Industrial Co., Ltd.
LandOfFree
Lombard effect compensation using a frequency shift does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Lombard effect compensation using a frequency shift, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Lombard effect compensation using a frequency shift will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1807472