SOUND RESTORING DEVICE AND SOUND RESTORING METHOD

摘要

A sound restoring device (101) for restoring restoration object sound in which a part of sound included in a mixed sound is missing comprises a mixed-sound separating section (103) for extracting the object sound contained in the mixed sound, a sound structure analyzing section (104) for creating at least one of phoneme series, character series, and note series in the portion where a part of sound of the extracted object sound is missing by using a sound structure knowledge database (105) where sound semantic contents are recorded, a same-sound characteristic domain analyzing section (106) for dividing the extracted object sound into time domain each having the same sound characteristic, a sound characteristic extracting section (107) for determining which time domain the missing part is positioned among the divided time domains and for extracting the sound characteristic of the object sound in the determined time domain, and a sound restoring section (108) for restoring the sound in the missing part of the object sound using the extracted sound characteristic and at least one of the phoneme series, character series, and note series created by the sound structure analyzing section (107).