1
学习音乐工具的提示词结构
AI音乐提示词需要五个要素：流派、情绪、乐器、节奏和参考。缺少其中任何一个都会产生普通的结果。
Example prompt
I want to generate music for [USE CASE: background, intro, ad, game, etc.]. Help me build a complete prompt. Ask me: 1) Genre and subgenre, 2) Mood and emotion, 3) Key instruments (3-5 max), 4) Tempo (BPM or relative: slow/mid/fast), 5) Any reference artists or tracks. Then assemble my answers into an optimized prompt for [SUNO/UDIO/MUSICGEN/etc.].
2
使用负面提示词和风格修饰符进行迭代
大多数AI音乐工具接受负面提示词。使用它们来删除您不想要的内容，并添加风格修饰符使结果更接近目标。
Example prompt
I generated this music track but it has these problems: [DESCRIBE ISSUES: too slow, wrong instruments, too busy, etc.]. Based on my original prompt: [PASTE ORIGINAL PROMPT]. Give me: 1) A revised positive prompt that addresses the issues, 2) Negative prompt keywords to exclude, 3) Three style modifier phrases I can add to shift the sound in the right direction.
3
将音乐与视频或内容时间匹配
为视频配乐时，使用AI将曲目分解成与场景变化相匹配的部分，而不仅仅是一个单一的氛围提示词。
Example prompt
I need background music for a [DURATION]-second video with these scene changes: [LIST SCENES WITH TIMESTAMPS]. For each scene, describe: the emotional tone needed, the instrumentation that fits, and the energy level (1-10). Then write a single music prompt that could work across all scenes, and identify where a version with a key change or energy shift should happen.
Ready to try Kaiber?
Follow this playbook with the actual tool