Background: Tinnitus is the perception of sound heard by a person without an external stimulus. About 10-14.5% of the world's population experience tinnitus, with the most common age group being 40-60. The management of tinnitus is still a challenge for researchers. Until now, there has been no effective management of tinnitus. The last therapy available is to make artificial sounds that imitate the tinnitus sound so that it can mask the actual tinnitus sound. In addition, music therapy is expected to reduce the stress level of tinnitus patients. Methods: This research is a follow-up study from previous research with 26 subjects. After getting the frequency and amplitude of tinnitus, a sound stimulus is made by combining tone variations and good generators obtained from audiometric masking. Tone variations are made using a basic tone that matches the patient's tinnitus frequency. After that, the patient will assess the results of the merger. Then all the participants will listen to dan evaluate the mixing song. Results: Of all subjective tinnitus sound generators, 3-5% have frequencies below 20 Hz and above 20,000 Hz. Songs made by recording in a soundproof room also have a frequency of 20 Hz and above 20,000 Hz, about 5 – 7%, where the song has an unnecessary frequency. The subjective tinnitus sound generator with the created song has an amplitude of < 85dB so that it can be adjusted with the volume control on the speaker. Conclusion: Tinnitus songs were made with the addition of Piano and Violin and Flute music, and Guitar, Bass and Balinese gamelan had a frequency of 20 – 20,000 Hz with an amplitude of < 85 dB, which can be adjusted with the volume so that it is safe and comfortable for tinnitus sufferers.