基于 SOC 职业分类
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。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/jmagly/aiwg --skill audio-extraction命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
WCAG accessibility analysis for color palettes including contrast ratios, compliance checking, and remediation suggestions. Use when user needs to verify colors meet accessibility standards.
Generate, analyze, compare, export, and suggest color palettes using color theory. Use when user asks about colors, palettes, color schemes, or needs help choosing colors for a project.
Research current color trends from Pantone, architecture, film, and design. Use when user asks about trending colors, popular palettes, or wants research-backed color inspiration.
| namespace | aiwg |
| platforms | ["all"] |
| name | audio-extraction |
| description | ffmpeg patterns for extracting audio from video files and transcoding between formats |
| category | media-curator |
Patterns for extracting audio tracks from video files and transcoding between audio formats using ffmpeg.
# Standard quality (128k) - good balance of size and quality
ffmpeg -y -i input.mp4 -vn -acodec libopus -b:a 128k output.opus
# High quality (192k) - for studio sessions and interviews
ffmpeg -y -i input.mp4 -vn -acodec libopus -b:a 192k output.opus
# Lower quality (96k) - for long-form content like full concerts
ffmpeg -y -i input.mp4 -vn -acodec libopus -b:a 96k output.opus
ffmpeg -y -i input.mp4 -vn -acodec flac output.flac
# VBR quality 2 (~190kbps, excellent quality)
ffmpeg -y -i input.mp4 -vn -acodec libmp3lame -q:a 2 output.mp3
# CBR 320kbps (maximum quality)
ffmpeg -y -i input.mp4 -vn -acodec libmp3lame -b:a 320k output.mp3
for f in video/*.mp4; do
[ -f "$f" ] || continue
base=$(basename "${f%.mp4}")
[ -f "audio/${base}.opus" ] && continue # Skip existing
ffmpeg -y -i "$f" -vn -acodec libopus -b:a 128k "audio/${base}.opus"
done
find video/ -name "*.mp4" -print0 | \
xargs -0 -P 4 -I{} sh -c '
base=$(basename "${1%.mp4}")
[ -f "audio/${base}.opus" ] || \
ffmpeg -y -i "$1" -vn -acodec libopus -b:a 128k "audio/${base}.opus"
' _ {}
ffmpeg -y -i input.flac -acodec libopus -b:a 128k output.opus
ffmpeg -y -i input.flac -acodec libmp3lame -q:a 2 output.mp3
ffmpeg -y -i input.opus -acodec libmp3lame -q:a 2 output.mp3
ffmpeg -y -i input.m4a -acodec libopus -b:a 128k output.opus
ffprobe -v quiet -show_entries format=bit_rate \
-of default=noprint_wrappers=1:nokey=1 input.opus
ffprobe -v quiet -show_entries stream=codec_name,sample_rate,channels,bit_rate \
-of json input.opus
ffmpeg -i input.opus -af "volumedetect" -f null /dev/null 2>&1 | grep max_volume
# If max_volume > 0 dB, audio is clipping
ffprobe -v quiet -show_entries format=duration \
-of default=noprint_wrappers=1:nokey=1 input.opus
| Source Content | Extract To | Bitrate | Rationale |
|---|---|---|---|
| Music video (studio) | Opus | 128k | Good balance for music |
| Live performance | Opus | 128k | Crowd noise doesn't need more |
| Studio session | Opus | 192k | Preserve detail |
| Interview | Opus | 96k | Speech doesn't need high bitrate |
| Full concert (2+ hours) | Opus | 96k | Size management |
| Archival source (FLAC) | FLAC | Lossless | Preserve original quality |
# Create file list
for f in segments/*.opus; do
echo "file '$f'" >> filelist.txt
done
# Concatenate without re-encoding
ffmpeg -f concat -safe 0 -i filelist.txt -c copy output.opus
# Split at specific timestamps (from setlist)
ffmpeg -i concert.opus -ss 00:00:00 -to 00:04:30 -c copy "01 - Jumpsuit.opus"
ffmpeg -i concert.opus -ss 00:04:30 -to 00:08:15 -c copy "02 - Levitate.opus"
# Extract frame at 5 seconds
ffmpeg -y -ss 5 -i video.mp4 -vframes 1 -q:v 2 cover.jpg
# Extract frame as WebP (smaller)
ffmpeg -y -ss 5 -i video.mp4 -vframes 1 cover.webp