When you are working with sound, you will often encounter a situation where the background music seems to hide or mask most of your dialog. In these situations what I recommend and like to do is open up a frequency analyzer. Hopefully one is available to you within your recording application. It is always a good idea to become familiar with the frequency range within the various tracks of your project so your mix comes out to your liking.
The next thing I do is check the frequency range of the person talking. For example, an average male frequency range can go from 500-5000k. Solo or just select your speaking track and watch the analyzer. Take note of the frequency range that is presented. Now go to your music track and play it. Since music if often more broad and covers more of the frequency gamut than dialog, you will see a much wider range here. Notice the range at where your speaker’s dialog fell and pull these frequencies down in the channel EQ. You can then bump up the outer range frequencies (not covered by your speaker) to give more gain or volume to those frequencies. What this will do is take the frequency range that your speaker falls down in volume while allowing the highs and lows of the music to stay up.
Essentially you will end up with music and dialog at close to the same volume without important frequencies being canceled out. Music is very important to the mood or feel of a scene and you really want it to do its job. So by keeping the volume where you like it and pulling out the shared frequencies, you achieve an overall clarity in your sound.
This technique is also very useful in song recording where you might have strings that have certain chords or notes that fall in the same range as your singer’s melody line.
–tim smith