摘要 |
In the present invention, a frequency domain conversion unit (11) divides an inputted signal into signals for each predetermined frame to generate a signal X (f, τ) for each first frequency division unit. A noise estimation signal generation unit (12) generates a signal Y (f, τ) for each second frequency division unit which is wider than the each of the first frequency division units. A peak range detection unit (15) obtains the peak range of the signal X (f, τ). A storage unit (13) stores the signals Y (f, τ). A signal comparison unit (14) calculates a representative value for each of the second frequency division units on the basis of the signal Y (f, τ) stored in the storage unit (13), and compares the representative value with the signal Y (f, τ) for each of the second frequency division units. A mask generation unit (16) generates a mask M (f, τ) for determining the degree of enhancement or suppression for each of the first frequency division units on the basis of the peak range and the comparison result obtained by the signal comparison unit (14). A mask application unit (17) multiplies the signal X (f, τ) by the mask M (f, τ). |