A portable Karaoke device is disclosed. An MP3 decoder and/or a communication interface is provided to download an accompaniment of a desired tune from a server and an expandable sound module and download memory is provided to store the accompaniment downloaded by the MP3 decoder and/or communication interface. A video signal received from a video unit by radio or an internally stored video signal is utilized as a background picture. A video signal mixed with a lyrics caption is transmitted by radio to the video unit to that the user can view the background picture and lyrics. In addition to a main microphone over which the user sings while listening to an accompaniment sound, there is provided an auxiliary microphone for transmitting a signal by radio or through the communication interface. The accompaniment sound is created in such a manner that maximum thirty-tow among one-hundred- twenty-eight musical instrument sounds are simultaneously mixed. While the accompaniment sound is outputted as an mixed musical instrument sound, it is modulated in such a manner that the phase and harmony thereof are varied into a desired phase and harmony. Accordingly, with no necessity for replacing a sound memory in which lyrics and accompaniments of newest songs are previously recorded and a video memory in which background pictures are stored, the user can simply and efficiently use the present system. Furthermore, a plurality of users can utilize the portable Karaoke device of the invention simultaneously. Moreover, the user's song is stored and then listened later or transmitted through the Internet in the form of music letter.