SPEECH PROCESSING APPARATUS, SPEECH PROCESSING METHOD, AND COMPUTER PROGRAM FOR SPEECH PROCESSING

摘要

PROBLEM TO BE SOLVED: To provide a speech processing apparatus that can appropriately set the gain of a speech signal according to the distance to a sound source to be an object of sound collection.SOLUTION: A speech processing apparatus 1 includes: a first distance calculation part (11) that calculates a ratio of power or amplitude between a first speech signal created by a sound collection part (2-1) and a human vibration signal created by a human vibration acquisition part (3), and calculates a first distance from the sound collection part (2-1) to the sound source according to the ratio and distance estimation information; a second distance calculation part (14) that calculates, for each frequency, a second distance from the sound collection part (2-1) to a sound source that has emitted a component of the frequency of a first frequency signal; and a gain determination part (15) that determines, for each frequency, a gain on the basis of a result of the comparison between the first distance and second distance.SELECTED DRAWING: Figure 3