Pitch detection method and apparatus uses voiced/unvoiced decision in a frame other than the current frame of a speech signal

• US 6,012,023 A
• Filed: 09/11/1997
• Issued: 01/04/2000
• Est. Priority Date: 09/27/1996
• Status: Expired due to Term

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Nishiguchi, Masayuki, Matsumoto, Jun, Iijima, Kazuyuki
• Primary Examiner(s): Dorvil, Richemond
• Application Number:
  US08/927,823
• Time in Patent Office: 845 Days
• Field of Search: 704/205, 704/206, 704/207, 704/208, 704/220, 704/223, 704/228
• US Class Current: 704/207
• CPC Class Codes: G10L 19/0212 using orthogonal transformation
• CPC Class Codes: G10L 19/09 Long term prediction, i.e. removing periodical redundancies, e.g. by using adaptive codebook or pitch predictor
• CPC Class Codes: G10L 25/90 Pitch determination of speech signals

Abstract

For realizing high-precision pitch detection even for speech signals in which half-pitch or double-pitch exhibits stronger autocorrelation than the pitch to be detected, an input speech signal is judged as to voicedness or unvoicedness and a voiced portion and an unvoiced portion of the input speech signal are encoded by a sinusoidal analytic encoding unit 114 and by a code excitation encoding unit 120, respectively, for producing respective encoded outputs. The sinusoidal analytic encoding unit 114 performs pitch search on the encoded outputs for finding the pitch information from the input speech signal and sets the high-reliability pitch information based on the detected pitch information. The results of pitch detection are determined using the high-reliability pitch information and the results of decision voicedness/unvoicedness of the frames other than the current frame.

Claims

1. A pitch detection method in an encoding method in which an input speech signal is divided on a time axis in terms of a pre-set frame and in which the frame-based speech signal is judged as to voiced/unvoiced, comprising:

a pitch searching step of detecting a pitch information under a pre-set pitch detection condition; and

a pitch determining step of determining a pitch of the current frame of the input speech signal based on the results of voiced/unvoiced decisions of the frames of the inputted speech signal other than the current frame on the time axis.

2. The pitch detection method as claimed in claim 1, wherein in the pitch searching step of detecting the pitch information under the pre-set pitch detecting condition, the pitch of the current frame of the input speech signal is determined using, as a parameter, the results of decisions of voiced/unvoiced of the input speech signal of past frames of the input speech signal on the time axis.

3. The pitch detection method as claimed in claim 1, further comprising a selecting step of using the voiced/unvoiced decisions of the input speech signal of the frames other than the current frame on the time axis for selecting whether the pitch information detected from the past frame is used as information for determining the final pitch for the current frame.

4. A speech signal encoding method in which an input speech signal is divided in terms of frame on a time axis and encoded on the frame basis, comprising:

a step of detecting a pitch of the input speech signal;

a predictive encoding step for finding short-term prediction residuals of the input speech signal;

a sinusoidal analysis encoding step for performing sinusoidal analysis encoding on the short-term prediction residuals found in the predictive encoding step;

a waveform encoding step for waveform encoding the input speech signal; and

a decision step for judging voiced/unvoiced of the input speech signal on the frame basis, wherein the pitch of the input speech signal of the current frame is determined also using the results of the voiced/unvoiced decision of the inputted speech signal of the frames other than the current frame on the time axis.

5. The speech encoding method as claimed in claim 4, wherein an encoded speech obtained by said sinusoidal analysis encoding step is outputted for the frame found to be voiced in the decision step, and wherein an encoded speech obtained by said waveform encoding step is outputted for the frame found to be unvoiced.

6. A speech signal encoding apparatus in which an input speech signal is divided in terms of frames on a time axis and encoded on the frame basis, comprising:

means for detecting a pitch of the input speech signal;

predictive encoding means for finding short-term prediction residuals of the input speech signal;

sinusoidal analysis encoding means for performing sinusoidal analysis encoding on the short-term prediction residuals found by said predictive encoding means;

waveform encoding means for waveform encoding the input speech signal; and

decision means for judging voiced/unvoiced of the input speech signal on the frame basis, wherein a pitch of the input speech signal of the current frame is determined using the results of the voiced/unvoiced decision of the inputted speech signal of the frames other than the current frame on the time axis.

7. The speech signal encoding apparatus as claimed in claim 6, wherein an encoded speech by said sinusoidal analysis encoding means is outputted for the frame found to be voiced by the decision means, and wherein an encoded speech by said waveform encoding means is outputted for the frame found to be unvoiced by the decision means.