菜单
Extract, enhance, and process audio from video files. Includes noise reduction, speech enhancement, and format conversion.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Calculate Four Pillars (四柱) from birth date/time using lunar calendar conversion. Foundation skill for all BaZi analysis.
Create and manage artifact.json for task outputs. Use when creating code files, documentation, or any files that should be tracked as artifacts.
Create and manage artifact.json for task outputs. Use at the START of every task to define planned deliverables before beginning research.
Provide career guidance and recommendations based on BaZi elemental strengths, Day Master traits, and current luck cycles.
Calculate and interpret 10-year luck cycles (大运) and annual fortune (流年) for timing insights and life trend predictions.
Analyze BaZi compatibility for romantic relationships (合婚). Examines elemental harmony, Day Master interaction, and long-term compatibility potential.
| name | audio-extractor |
| description | Extract, enhance, and process audio from video files. Includes noise reduction, speech enhancement, and format conversion. |
This skill extracts and enhances audio from video files for processing or standalone use.
Extract audio from: ./output/downloads/VIDEO_ID.mp4
Extract and enhance audio from VIDEO_ID
Extract audio as WAV for transcription
ffprobe -v quiet -print_format json -show_streams \
-select_streams a:0 \
"./output/downloads/VIDEO_ID.mp4"
To MP3 (compressed, good for sharing):
ffmpeg -i "./output/downloads/VIDEO_ID.mp4" \
-vn -acodec libmp3lame -q:a 2 \
"./output/downloads/VIDEO_ID.mp3"
To WAV (uncompressed, best for processing):
ffmpeg -i "./output/downloads/VIDEO_ID.mp4" \
-vn -acodec pcm_s16le -ar 44100 \
"./temp/VIDEO_ID.wav"
To AAC (high quality, smaller):
ffmpeg -i "./output/downloads/VIDEO_ID.mp4" \
-vn -acodec aac -b:a 192k \
"./output/downloads/VIDEO_ID.m4a"
Normalize Volume:
ffmpeg -i "./temp/VIDEO_ID.wav" \
-af "loudnorm=I=-16:TP=-1.5:LRA=11" \
"./output/downloads/VIDEO_ID_normalized.mp3"
Reduce Noise + Normalize:
ffmpeg -i "./temp/VIDEO_ID.wav" \
-af "highpass=f=80,lowpass=f=12000,loudnorm=I=-16:TP=-1.5:LRA=11" \
"./output/downloads/VIDEO_ID_enhanced.mp3"
Speech Enhancement (for voice clarity):
ffmpeg -i "./temp/VIDEO_ID.wav" \
-af "highpass=f=100,lowpass=f=8000,loudnorm,compand=attacks=0:points=-80/-80|-45/-15|-27/-9|0/-7|20/-7:gain=5" \
"./output/downloads/VIDEO_ID_speech.mp3"
## ✅ Audio Extracted
**Source**: VIDEO_ID.mp4
**Output**: VIDEO_ID.mp3
### Audio Info
- **Duration**: 10:32
- **Sample Rate**: 44.1 kHz
- **Channels**: Stereo
- **Bitrate**: 192 kbps
- **Size**: 12.5 MB
### Quality Assessment
- ✅ Speech clarity: Good
- ✅ Background noise: Minimal
- ✅ Volume levels: Normalized
**Next step**: Use `speech-transcriber` to convert to text
./output/downloads/{video_id}.mp3./output/downloads/{video_id}_enhanced.mp3./temp/{video_id}.wav