To perform high-grade sound processing by converting the pitch of inputted sounds on the basis of performance information and outputting converted sounds together with inputted sounds.

A sound processing device is provided with a microphone 1 through which sounds are inputted, an external storage device 13 where accompaniment sound information which is used to obtain harmony sounds in accordance with progress of a musical piece is previously stored, and a pitch conversion LSI 15 which converts direct sounds inputted from the microphone 1 to harmony sounds on the basis of performance information read out from the external storage device 13. The external storage device 13 is, for example, a KARAOKE (orchestration without lyrics) device. When the KARAOKE device 13 reads out musical piece data and inputs it to the sound processing device and direct sounds are inputted from the microphone 1, the pitch conversion LSI 15 generates harmony sounds from direct sounds on the basis of harmony information in performance information, and a mixer 6 synthesizes and outputs harmony sounds and direct sounds.