A method and device for producing ring tones in high polyphony in real-time in a network component such as a mobile terminal or a server of a service provider. The network component has one or more ring tones in MIDI files of high polyphony, and a MIDI player of lower polyphony for producing sounds from the scaled down version of the MIDI files in real-time. In order to produce sounds indicative of the high-polyphony ring tones in real-time, the high-polyphony MIDI files are converted to compressed files in a non real-time manner. The converted files are stored in a storage so as to allow a compressed file player (such as Truetone, MP3, wav, AAC, RealAudio, Vorbis) to produce sounds from the converted files. A file lock is provided to the converted files so that they cannot be forwarded, thereby protecting the copyrights of the ring tone composer.