# AI Music Generation Prompting — a skill for scoring film in sections
Source: ekly.ai/guides/ai-music-prompting — free to use with any AI assistant.

## Hard rules
1. Match the prompting style to the engine: free-text composer brief vs. tag list — don't swap them.
2. Vocals are often on by default — set the instrumental toggle; don't rely on prose alone under dialogue.
3. Timestamped section briefs steer arrangement direction; they are not frame-accurate contracts.
4. Score a film as separate cues per emotional section, not one flat bed under everything.
5. Keep shared production vocabulary across cues; vary only energy and arrangement.
6. Mix music ~12–15 dB under dialogue (15–20 dB for comedy); duck under lines; leave midrange room in the brief.
7. Score the ad's genre dials — don't generate a music-video bed for "energy." Name the bed's job per section.
8. Key the ducker off dialogue shifted ~150 ms earlier (look-ahead). On a hard cut into a line, a reactive-only ducker lets music peek over the first syllable. When a QC gate fails on a boundary, fix the mix — never relax the threshold.
9. If music must read from frame 0 under VO, re-lock mix policy (parallel bed vs dialogue-led) and verify open-window RMS on the first ~2s; trim quiet intros that start after picture.

## Related
- ElevenLabs music section — ekly.ai/guides/elevenlabs-prompting#music
- ElevenLabs in Ekly — ekly.ai/models/elevenlabs