摘要 |
A model sound index value calculating means (123) calculates, according to a prescribed calculation formula, a model sound index value which is an index value of the maximum value of power for each frequency band of the model sound which is a model of a target sound. A source sound index value calculating means (124) calculates, according to a prescribed calculation formula, a source sound index value which is an index value of power for each frequency band with respect to each of frames extracted by a predetermined time length from a source sound signal used for generating a masker sound signal. A masking performance calculating means (125) calculates a performance index value which is an index value of performance of masking the model sound by a sound represented by a block formed of a predetermined number of consecutive frames extracted from the source sound signal, by using the model sound index value and the source sound index value. A frame selecting means (126) determines a block to be used for generating the masker sound based on the performance index value. |