Narration BGM is one of the easiest places to overdo music. A dramatic cue can make a sentence feel important, but it can also make the speaker harder to understand. The music should leave room in the midrange and avoid sudden melodic events under key phrases.
Low density is not the same as low quality. A simple pad, soft pulse, light piano, or gentle texture can make a voice feel warmer without competing. Harsh snares, busy arpeggios, and vocal chops are risky because they pull attention away from language.
The prompt should include speech context: documentary narration, tutorial, interview intro, meditation guide, product explainer, or news recap. Each needs a different level of motion. A meditation guide might need almost no rhythm, while a tech explainer can use a subtle pulse.
If the video or podcast is multilingual, avoid lyrics entirely. Even background humming can make speech feel crowded. Instrumental mode should be treated as a safety default for narration-heavy content.
A useful prompt is: 'subtle narration background music for a calm product explainer, warm pad, soft piano accents, very light pulse, low midrange masking, no vocals, no humming, no busy lead melody.' The words protect the voice before the music is generated.