A mobile phone (100) detects at least a direction of an utterance position of a user from the voice input to a plurality of microphones (121, 122) and controls an output direction of a voice out of a speaker unit (110) to match the detected utterance position. Accordingly, it is possible to provide sound with sufficient volume only for a specific user without causing noise in the neighborhood. The microphones may form a matrix together with piezoelectric vibrators that form a speaker unit, or may be formed separately from the speaker unit.