Grab headphones, a music playing device, and a song of your choosing. Sit down in a quiet room in a chair where you will not be disturbed. Play the first 10 seconds of the song and then pause it. Close your eyes and try to imagine the first 10 seconds which you just played. Record and number any breaks in focus. Continue with this process until the song is over. Gradually increase the amount of time. For example, once you find that imagining 10 seconds is easy, try imaging 20 seconds, then 30 at a time.
Do not proceed to the next exercise until you can make it through a whole song (5+ minutes) without more than 3 breaks in focus. You should be able to vividly imagine the song before proceeding with the next exercise.
This practice should take some time to get right. Due to this and the nature of the exercise, every day you should choose a different song.