As you've discovered, the dialogue emphasis feature isn't meant for music. With certain programming which includes music, there is a tradeoff.
When you look at HA's, models with adjustable sound tuning (bass, treble & midrange) can compensate for distortion caused by the dialogue emphasis...