声音制作规范,Jiuge_Flow_Perfect_V1.skill

# SKILL.md - 九歌完美版音频处理流水线

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "声音制作规范,Jiuge_Flow_Perfect_V1.skill" with this command: npx skills add lzgsrbs/jiuge-base-behavior

SKILL.md - 九歌完美版音频处理流水线

名称

Jiuge Flow Perfect V1 - 九歌音频大师完美版

版本

V1.2.0

描述

基于2026-05-01 21:47《敬兄弟》no-wash完美处理经验封装的音频处理Skill。执行"01_Sep → 03_Dry → 04_Master"三步流程,跳过洗声节省25%时间,全链路48kHz,输出320kbps MP3预览版。

核心特性

  • 默认no-wash:智能跳过洗声,节省约25%处理时间
  • 全链路48kHz:从分离到母带保持统一采样率
  • 双版本输出:320kbps母带 + 128kbps预览版
  • 自动归档:按"项目名+时间戳"自动归档到Jiuge_Audio_Projects
  • 断点续处理:支持从任意步骤中断点继续

执行流程

第一步:人声分离 (01_Sep)

  • 引擎:Kim_Vocal_2.onnx (MDX-Net)
  • 模型路径:/Applications/Ultimate Vocal Remover.app/Contents/Resources/models/MDX_Net_Models
  • 输出格式:WAV 48kHz
  • 输出文件:(Vocals)_Kim_Vocal_2.wav, (Instrumental)_Kim_Vocal_2.wav

第二步:去混响脱水 (03_Dry)

  • 引擎:UVR-De-Echo-Aggressive.pth (VR Architecture)
  • 模型路径:/Applications/Ultimate Vocal Remover.app/Contents/Resources/models/VR_Models
  • 输入:第一步的Vocals文件
  • 输出:(No Echo)_UVR-De-Echo-Aggressive.wav

第三步:母带处理 (04_Master)

  • 转换工具:ffmpeg
  • 母带版本:320kbps MP3, 48kHz, 立体声
  • 预览版本:128kbps MP3, 48kHz, 立体声
  • 命名规则:{项目名}_Master.mp3, {项目名}_Preview.mp3

使用方法

基本用法(默认no-wash)

~/Desktop/Jiuge_Flow_Perfect_V1.skill/scripts/jiuge_flow.sh "文件名.mp3" [模式编号]

强制wash模式

~/Desktop/Jiuge_Flow_Perfect_V1.skill/scripts/jiuge_flow.sh "文件名.mp3" 1 wash

参数说明

  • $1:输入文件名(必填)
  • $2:处理模式(可选,默认1)
    • 1:民歌高亢版(默认)
    • 2:美声共鸣版
    • 3:流行中度脱水
    • 4:会议轻度脱水
    • 5:大教堂重度脱水
  • $3:wash开关(可选,默认no-wash)
    • wash:执行完整四步流程(含洗声)
    • no-wash:跳过洗声,三步完成

技术规范

硬性指标

项目规格
采样率48kHz(全链路统一)
母带码率320kbps MP3
预览码率128kbps MP3
声道立体声
输出格式MP3

模型配置

步骤模型架构用途
分离Kim_Vocal_2.onnxMDX-Net人声/伴奏分离
脱水UVR-De-Echo-Aggressive.pthVR去混响

环境要求

  • Python:3.11+(uvr_auto conda环境)
  • audio-separator:0.24.1+
  • ffmpeg:4.3.2+
  • 模型文件:已安装UVR5模型

修复记录

V29 Bug修复

  • 问题:V29脚本dry_cmd为None导致TypeError
  • 原因:自动化逻辑中变量未正确初始化
  • 解决:按21:47手动成功流程重新编写分步脚本
  • 验证:2026-05-01 21:47《敬兄弟》处理成功

作者

九歌传媒 AI助手

更新日志

  • V1.2.0 (2026-05-01 23:51): 修复03_Dry文件名选择bug,排除Instrumental文件,确保选择主唱人声
  • V1.1.0 (2026-05-01 22:47): 修复洗声逻辑,补充完整的背景合唱分离功能
  • V1.0.0 (2026-05-01): 基于21:47完美处理经验首次封装

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

Prompt Wizard

Generate high-quality English prompts for ChatGPT Image 2. Use when user wants to create AI image prompts, needs GPT-Image-2 prompt writing help, describes a...

Registry SourceRecently Updated
General

App Builder

Full-stack application builder that creates web apps, APIs, mobile apps, and more from natural language requests. Use when the user wants to build a new appl...

Registry SourceRecently Updated
General

Golang Testing

Provides a comprehensive guide for writing production-ready Golang tests. Covers table-driven tests, test suites with testify, mocks, unit tests, integration...

Registry SourceRecently Updated
2280samber
General

Infinitecampus Mcp

This skill should be used when the user asks about Infinite Campus (Campus Parent) data for their student(s). Triggers on phrases like "check grades", "what'...

Registry SourceRecently Updated