Consonant-segment detection apparatus and consonant-segment detection method

摘要

A signal portion is extracted from an input signal for each frame having a specific duration to generate a per-frame input signal. The per-frame input signal in a time domain is converted into a per-frame input signal in a frequency domain, thereby generating a spectral pattern. Subband average energy is derived in each of subbands adjacent one another in the spectral pattern. The subband average energy is compared in at least one subband pair of a first subband and a second subband that is a higher frequency band than the first subband, the first and second subbands being consecutive subbands in the spectral pattern. It is determined that the per-frame input signal includes a consonant segment if the subband average energy of the second subband is higher than the subband average energy of the first subband.

主权项

1. A consonant-segment detection method comprising the steps of:
extracting a signal portion from an input signal for each frame having a specific duration to generate a per-frame input signal; converting the per-frame input signal in a time domain into a per-frame input signal in a frequency domain, thereby generating a spectral pattern; deriving subband average energy in each of subbands adjacent one another in the spectral pattern; comparing the subband average energy in at least one subband pair of a first subband and a second subband that is a higher frequency band than the first subband, the first and second subbands being consecutive subbands in the spectral pattern; and determining that the per-frame input signal includes a consonant segment if a positive result of comparison is obtained, the positive result indicating that the subband average energy of the second subband is higher than the subband average energy of the first subband, wherein the determining step has a first determining step and a second determining step, wherein the first determining step is performed for counting the number of a plurality of subband pairs of the first and second subbands if the positive result is obtained for the subband pairs and determining that the per-frame input signal includes the consonant segment if the counted number is equal to or larger than a predetermined first threshold value, and the second determining step is performed for counting the number of a plurality subband pairs of the first and second subbands with weighting if the positive result is obtained for the subband pairs, the subband pairs being consecutive subband pairs in the spectral pattern and a subband of higher frequency in each former subband pair being a subband of lower frequency in each latter subband pair that follows each former subband pair in the consecutive subbands, and determining that the per-frame input signal includes the consonant segment if the weighted counted number is equal to or larger than a predetermined second threshold value, wherein the consonant-segment detection method further comprises the steps of: deriving a noise level of the per-frame input signal; and selecting the first determining step if the noise level is smaller than a predetermined fourth value, and selecting the second determining step if the noise level is equal to or larger than the predetermined fourth value.