mcpField
M
music-video-subtitle-generator

music-video-subtitle-generator

github.com/sjh00Creative & media4054 installsUpdated 9w ago
Repository

About music-video-subtitle-generator

music-video-subtitle-generator is a skill for writing multi-shot typography prompt scripts for music videos, intended for use with H3. It structures text as part of the visual space, synchronizes lyrics with on-screen typography, and specifies shot transitions.

Key features

  1. 01Define aspect ratio, duration, and music window
  2. 02Lock lyrics or original short text
  3. 03Separate typography, character, and scene guides
  4. 04Write timed multi-shot prompt scripts
  5. 05Specify beat-aware cuts and transitions

Use cases

  1. 01Planning typography-led music video shots
  2. 02Synchronizing vocal lines and displayed text
  3. 03Preparing prompts for H3 video generation

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/sjh00/minimax-h3-storyboard-prompt-skills --skill music-video-subtitle-generator

After installing, type/music-video-subtitle-generatorin Claude Code to confirm it loaded.

Bundled filesmeta.yaml

Details

From the project README

音乐 MV 贴字 · 提示词写法

交付:多分镜 Prompt 脚本(可直接按镜投 H3)。拼接只写设计衔接句,不写工程步骤。

原则

  • 文字是空间设计层,非底端字幕条(除非用户要普通字幕)
  • 用户音乐=主床;>15s 拆 2–5s 多镜
  • 有人声时 Typography 与表演词逐字一致;每镜一个主文字事件
  • 默认硬切跟鼓点;禁无意义淡入淡出

写前锁定

画幅(9:16/16:9/1:1/21:9)· 目标时长 · 音乐窗口 · 锁定歌词(用户词不改;无词则先锁原创短词)

三卡(写入 Global Lock)

卡只写勿写
文字卡字体、质感、动效语言人物场景
人物卡脸气质服装姿态复杂场景抢戏
场景卡空间光影胶片感大段可读字

分镜脚本模板

[Global Aesthetic & Character Lock]
画幅 / 风格颗粒与色 / 三卡引用 / 禁淡入淡出、禁油亮AI脸 …

Shot 1 (0.0s-3.5s)
Vocal Line: "…"
Typography: "…"   # 空间位置:勿挡眼/主表情
Visual & Action: …
Camera & Motion: …
Transition Out: bass-hit hard cut / match pan …

Shot 2 …

每镜可再压成 h3-prompt-writing 字段;对白/歌词进 <d> 保留原文。

衔接句(写在脚本末)

  1. 切点:句间停顿或强鼓点
  2. 同场景:尾帧→下镜首帧续写
  3. 跨景:同向运镜或 Match Cut
  4. 文字:切点砸出/砸入跟 Bass
  5. 全局同一颗粒/光向描述

自检

  • >15s 已拆镜
  • 词与贴字一致
  • 三卡不串味
  • 硬切与拍点意图写清

Comments

Markdown supported