| 摘要 |
PROBLEM TO BE SOLVED: To extract or eliminate a stationary sound in real time from an audio signal in which a plurality of sound sources are included.SOLUTION: An audio signal processing device comprises: a short-time Fourier transform unit 4 for generating a frequency-domain signal in which short-time Fourier transform is applied to an inputted audio signal; a stationary sound determination unit 5 for determining whether the waveform of a peak portion included in the signal waveform of the frequency domain is a stationary sound or not; a filter coefficient arithmetic unit 6 for dynamically calculating a filter coefficient on the basis of the result of determination by the stationary sound determination unit 5; a comb-shaped filter unit 7, driven by the filter coefficient calculated by the filter coefficient arithmetic unit 6, for filtering the frequency-domain signal; and an inverse Fourier transform unit 8 for converting the output of the comb-shaped filter unit 7 into a time-domain signal and outputting the converted signal.SELECTED DRAWING: Figure 7 |
