Data processing: speech signal processing – linguistics – language – Speech signal processing – For storage or transmission
Reexamination Certificate
1997-09-11
2001-06-05
Korzuch, William R. (Department: 2641)
Data processing: speech signal processing, linguistics, language
Speech signal processing
For storage or transmission
C704S217000
Reexamination Certificate
active
06243672
ABSTRACT:
BACKGROUND OF THE INVENTION
Field of the Invention
This invention relates to a speech encoding method and apparatus in which an input speech signal is split on the time axis in terms of a pre-set block as an encoding unit and encoded from one such encoding unit to another. The invention also relates to a pitch detection method employing the speech encoding method and apparatus.
Description of the Related Art
Up to now, there are known a variety of encoding methods for performing signal compression by exploiting statistic properties in the time domain and frequency domain of audio signals, inclusive of speech and acoustic signals, and psychoacoustic properties of the human being. These encoding method are roughly classified into encoding in the time domain, encoding in the frequency domain and analysis-synthesis encoding.
Among the techniques for high-efficiency encoding of speech signals are sinusoidal analysis encoding, such as harmonic encoding or multi-band excitation (MBE) encoding, sub-band coding (SBC), linear predictive coding (LPC), discrete cosine transform (DCT), modified DCT (MDCT) and fast Fourier transform (FFT).
Meanwhile, in the sinusoidal synthetic encoding generating excitation signals using the pitch of an input speech signal as a parameter, pitch detection plays an important role. With a pitch detection method employing an autocorrelation method used in a conventional speech signal encoding circuit and which is seasoned with a fractional search with the sample shifting of not more than one sample for improving pitch detection accuracy, if the half-pitch or the double pitch exhibits stronger correlation than the pitch desired to be detected in the speech signal, pitch detection tends to result in failure.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide a pitch detection method capable of correctly detecting the pitch for a speech signal in which the half-pitch or the double pitch exhibits stronger correlation than the pitch desired to be detected in the speech signal.
It is another object of the present invention to provide a speech signal encoding method and apparatus capable of producing a highly clear natural playback speech devoid of extraneous noise by application of the above-mentioned pitch detection method.
The present invention provides a pitch detection method in which an input speech signal is divided on the time axis in terms of pre-set encoding units for detecting the pitch corresponding to the fundamental period of the encoding-unit-based speech signal. The method includes a pitch searching step of detecting the pitch information under a pre-set pitch detection condition, a step of setting the high-reliability pitch information satisfying a condition which becomes true if the likeliness to pitch is higher than under the pitch detecting condition based on the detected pitch information, speech level of the input speech signal and on the autocorrelation peak value of the input speech signal, and a step of determining the pitch based on the set high-reliability pitch information.
The pitch detection method according to the present invention permits high-precision pitch detection without mistaken detection of the half-pitch or the double-pitch.
The present invention also provides a speech signal encoding method and device in which an input speech signal is divided on the time axis in terms of a pre-set encoding unit and encoded on the encoding unit basis. The encoding method and device include predicting encoding of detecting the pitch by the above-defined pitch detection method for finding short-term prediction residuals of the input speech signals, sinusoidal analytic encoding of performing sinusoidal analytic encoding on the short-term prediction residuals thus found, waveform analysis encoding of performing waveform analysis encoding on the input speech signals and judgment of the voicedness/unvoicedness of the input speech signals.
With the above-described speech encoding method and device according to the present invention, pitch detection can be performed without mistaken detection of the half-pitch or double-pitch in the speech signals, thus enabling explosives or fricatives, such as p, k or t, to be reproduced clearly, while there is produced no extraneous sound in a transition portion between the voiced and unvoiced portions, thus assuring reproduction of the clear natural speech devoid of buzzing.
REFERENCES:
patent: 4809334 (1989-02-01), Bhaskar
patent: 5226108 (1993-07-01), Hardwick et al.
patent: 5699477 (1997-12-01), McCree
patent: 5704000 (1997-12-01), Swaminathan et al.
patent: 5745871 (1998-04-01), Chen
patent: 5765125 (1998-06-01), Daugherty et al.
patent: 5774836 (1998-06-01), Bartkowiak et al.
patent: 5826222 (1998-10-01), Griffin
patent: 0266868 (1988-05-01), None
patent: 0280827 (1988-09-01), None
patent: 0333121 (1989-09-01), None
Iijima Kazuyuki
Matsumoto Jun
Nishiguchi Masayuki
Korzuch William R.
Maioli Jay H.
Sony Corporation
Zintel Harold
LandOfFree
Speech encoding/decoding method and apparatus using a pitch... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Speech encoding/decoding method and apparatus using a pitch..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Speech encoding/decoding method and apparatus using a pitch... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2467987