メニュー
Audio engineering — mastering, mixing, EQ, compression, loudness standards, synthesis, podcast production, music theory, spectrum analysis.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/Tibsfox/gsd-skill-creator --skill audio-engineeringコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
Use the moment a request is underspecified and you are about to act on an ASSUMED goal. Enumerate the candidate goals the request could mean, project the action or artifact each one implies, and check whether they diverge: if two or more plausible goals lead to materially different actions or artifacts, ask exactly ONE targeted clarifying question; otherwise proceed on the most-likely goal and state the assumption in one line. Clarification is an evidence-producing action, not a delay — but one question, never an interrogation. Distinct from gsd-spec-phase (GSD-phase-bound, emits a heavy SPEC.md) and intent-router (fetch strategy, not goal disambiguation). Backed by agent goal-state inference (arxiv 2606.16813v1). Triggers on acting under an assumed goal when the request admits more than one materially different reading.
Use at write time to vector/embedding memory — Grove content-addressed insertion, chroma/pgvector upserts, memory-consolidation promoting session traces to MEMORY.md, or embedding externally-ingested content. Scores each candidate record against a fixed panel of sentinel queries and quarantines any record that would become the nearest neighbor of too many unrelated queries — a hub — whether from adversarial poisoning or accidental over-generality. This is the memory-record-side sibling of skill-injection-guardian (file-side) and the write-side complement of memory-use-warrant (read-side). Quarantine, never silently drop; a human reviews. Backed by the admission-time hubness gate (arxiv 2606.19692v1). Triggers on inserting into vector memory, consolidating memory, and embedding stranger content.
Run this appropriateness check the moment you are about to integrate a retrieved long-term memory — a Grove content-addressed hit, a chroma/pgvector neighbour, a memory-consolidation digest, or a MEMORY.md line — into a response, especially anything touching private origins, Fox Companies IP, credentials, or Center Camp / consent-governed content. It answers a question intent-router never asks: not WHAT to fetch or HOW, but WHETHER a correctly-retrieved item should reach output. Relevance is not appropriateness — a perfect similarity match can still be a boundary violation. Default is FAIL-CLOSED: if in-context authorization is uncertain, the memory may inform behaviour but must not be surfaced. Backed by RBI-Eval (arxiv 2606.06055v1). Triggers on surfacing recalled sensitive memory into a response.
SKILL.md を表示中
SOC 職業分類に基づく
| name | audio-engineering |
| description | Audio engineering — mastering, mixing, EQ, compression, loudness standards, synthesis, podcast production, music theory, spectrum analysis. |
| activationKeywords | ["audio","mastering","mixing","EQ","equalization","compressor","compression","limiter","loudness","LUFS","true peak","EBU","podcast","waveform","spectrum","synthesis","synthesizer","FM synthesis","subtractive","granular","modular","BPM","tempo","key detection","music theory","chord","scale","sox","ableton","DAW","reverb","delay","noise reduction","normalization","dithering","sample rate","bit depth"] |
| type | skill |
| category | media |
| status | stable |
| origin | tibsfox |
| modified | false |
| first_seen | "2026-04-11T00:00:00.000Z" |
| first_path | examples/skills/media/audio-engineering/SKILL.md |
| superseded_by | null |
Built on research from 30 Music cluster projects, 360 PNW musicians deep-dived (S36/SPS), Ableton Live research (ABL), Deep Audio (DAA), Dead Frequencies (DFQ), and High Fidelity amplifier analysis (HFR/HFE).
Expert-level audio engineering covering mastering, mixing, loudness standards, synthesis, podcast production, music theory, and spectrum analysis. Works alongside the ffmpeg-media skill for codec/format operations.
| Platform | Target LUFS | True Peak | Standard |
|---|---|---|---|
| Spotify | -14 LUFS | -1 dBTP | AES streaming |
| Apple Music | -16 LUFS | -1 dBTP | Sound Check |
| YouTube | -14 LUFS | -1 dBTP | ITU-R BS.1770 |
| Podcast (Apple) | -16 LUFS | -1 dBTP | Apple spec |
| Podcast (Spotify) | -14 LUFS | -1 dBTP | Spotify spec |
| Broadcast TV | -24 LUFS | -2 dBTP | EBU R128 |
| Broadcast US | -24 LKFS | -2 dBTP | ATSC A/85 |
| CD master | -9 to -12 LUFS | -0.3 dBTP | Red Book |
| Film/Cinema | -24 LUFS | -1 dBTP | SMPTE RP 200 |
# Measure integrated loudness (LUFS) with ffmpeg
ffmpeg -i input.wav -af loudnorm=print_format=json -f null - 2>&1 | grep -A20 "Parsed_loudnorm"
# Full EBU R128 scan
ffmpeg -i input.wav -af ebur128=peak=true -f null - 2>&1 | tail -20
# Loudness normalization to -14 LUFS (two-pass for accuracy)
# Pass 1: measure
ffmpeg -i input.wav -af loudnorm=I=-14:LRA=11:TP=-1:print_format=json -f null - 2>&1 > /tmp/loudnorm.json
# Pass 2: apply (use measured values from pass 1)
ffmpeg -i input.wav -af loudnorm=I=-14:LRA=11:TP=-1:measured_I=-18.5:measured_LRA=9.2:measured_TP=-0.5:measured_thresh=-28.3 output.wav
# Normalize peak to -1 dBFS
sox input.wav output.wav gain -n -1
# Compressor (threshold -20dB, ratio 4:1, attack 5ms, release 50ms)
sox input.wav output.wav compand 0.005,0.05 -20,-20,-10,-10,0,-6
# 3-band EQ (low shelf +3dB at 200Hz, mid cut -2dB at 2kHz, high shelf +1dB at 8kHz)
sox input.wav output.wav bass +3 200 equalizer 2000 1q -2 treble +1 8000
# Noise reduction (profile then reduce)
sox noisy.wav -n noiseprof /tmp/noise.prof
sox noisy.wav clean.wav noisered /tmp/noise.prof 0.21
# Generate tone (440Hz sine, 3 seconds)
sox -n -r 44100 -c 1 tone.wav synth 3 sine 440
# Spectrum analysis (generate spectrogram PNG)
sox input.wav -n spectrogram -o spectrum.png
Input → EQ (corrective) → Compression → EQ (tonal) → Stereo Width → Limiting → Dithering → Output
# Full mastering chain: EQ → compression → limiting → loudness normalization
ffmpeg -i mix.wav -af "\
equalizer=f=80:t=h:w=100:g=2,\
equalizer=f=3000:t=h:w=1000:g=-1.5,\
equalizer=f=12000:t=h:w=2000:g=1,\
acompressor=threshold=-18dB:ratio=3:attack=10:release=100:knee=6,\
alimiter=limit=-1dBFS:level=false,\
loudnorm=I=-14:LRA=11:TP=-1\
" -ar 44100 -sample_fmt s16 mastered.wav
# Dithering (16-bit with triangular dither for CD)
ffmpeg -i master_24bit.wav -af "dither=method=triangular" -sample_fmt s16 -ar 44100 cd_master.wav
| Band | Range | Character | Common Uses |
|---|---|---|---|
| Sub-bass | 20-60 Hz | Felt, not heard | Kick fundamental, sub bass |
| Bass | 60-250 Hz | Warmth, body | Bass guitar, kick punch, vocal warmth |
| Low-mid | 250-500 Hz | Muddiness zone | Cut here to clean up mixes |
| Mid | 500-2000 Hz | Body, presence | Vocal clarity, guitar body |
| Upper-mid | 2-4 kHz | Presence, bite | Vocal intelligibility, guitar attack |
| Presence | 4-6 kHz | Definition, edge | Consonant clarity, string attack |
| Brilliance | 6-12 kHz | Air, shimmer | Cymbals, vocal air, acoustic sparkle |
| Ultra-high | 12-20 kHz | Air, sparkle | Subtle sheen (careful: sibilance) |
| Source | Threshold | Ratio | Attack | Release | Knee |
|---|---|---|---|---|---|
| Vocals | -18 to -12 dB | 2:1 to 4:1 | 5-15 ms | 40-80 ms | Soft |
| Drums (bus) | -15 to -10 dB | 3:1 to 6:1 | 10-30 ms | 50-100 ms | Hard |
| Bass | -15 to -8 dB | 3:1 to 8:1 | 10-30 ms | 100-200 ms | Hard |
| Acoustic guitar | -20 to -12 dB | 2:1 to 4:1 | 10-25 ms | 100-150 ms | Soft |
| Mix bus | -20 to -15 dB | 1.5:1 to 2:1 | 10-30 ms | 100-300 ms | Soft |
| Podcast | -20 to -15 dB | 3:1 to 5:1 | 5-10 ms | 50-100 ms | Soft |
| Type | How It Works | Character | Classic Synths |
|---|---|---|---|
| Subtractive | Oscillator → Filter → Amplifier | Warm, analog, rich | Minimoog, Prophet-5, Juno-106 |
| FM | Operators modulating each other's frequency | Metallic, bell-like, bright | DX7, FM8 |
| Wavetable | Morphing between stored waveforms | Evolving, complex, modern | PPG Wave, Serum, Vital |
| Granular | Tiny audio grains layered and scattered | Atmospheric, textural, ambient | Granulator, Pigments |
| Additive | Sum of individual sine wave partials | Precise, organ-like | Kawai K5, Razor |
| Physical modeling | Mathematical model of physical instrument | Realistic, expressive | Chromaphone, Pianoteq |
| Sample-based | Recorded audio, pitch-shifted and layered | Realistic, natural | Kontakt, Sampler |
C
F G
Bb D
Eb A
Ab/E
| Name | Numerals | Example in C | Use |
|---|---|---|---|
| Pop | I-V-vi-IV | C-G-Am-F | 80% of pop music |
| Blues | I-IV-V | C-F-G | Blues, rock |
| Jazz ii-V-I | ii-V-I | Dm7-G7-Cmaj7 | Jazz standard |
| Andalusian | i-VII-VI-V | Am-G-F-E | Flamenco, dramatic |
| Canon | I-V-vi-iii-IV-I-IV-V | C-G-Am-Em-F-C-F-G | Pachelbel, ballads |
| Minor blues | i-iv-V | Am-Dm-E | Minor blues |
# Record from default mic (sox)
sox -d -r 44100 -c 1 -b 16 recording.wav
# Record with ffmpeg (specify ALSA device on Linux)
ffmpeg -f alsa -i default -ar 44100 -ac 1 recording.wav
# 1. Noise reduction
sox recording.wav -n trim 0 0.5 noiseprof /tmp/noise.prof
sox recording.wav clean.wav noisered /tmp/noise.prof 0.21
# 2. Normalize + compress + EQ for voice
ffmpeg -i clean.wav -af "\
highpass=f=80,\
lowpass=f=12000,\
equalizer=f=3000:t=h:w=1000:g=2,\
acompressor=threshold=-20dB:ratio=4:attack=5:release=50,\
loudnorm=I=-16:TP=-1\
" -ar 44100 podcast_ready.wav
# 3. Export MP3 for distribution
ffmpeg -i podcast_ready.wav -c:a libmp3lame -b:a 128k \
-metadata title="Episode Title" \
-metadata artist="Show Name" \
-metadata album="Podcast Name" \
-metadata genre="Podcast" \
episode.mp3
# 4. Generate waveform for show notes
ffmpeg -i episode.mp3 -filter_complex "showwavespic=s=1920x200:colors=0x1a1a2e" -frames:v 1 waveform.png
# Set all metadata
ffmpeg -i episode.mp3 -c copy \
-metadata title="EP 42: The Memory Architecture" \
-metadata artist="GSD Podcast" \
-metadata album="Getting Shit Done" \
-metadata track="42" \
-metadata date="2026" \
-metadata comment="LOD-tiered memory system deep dive" \
tagged.mp3
# Install aubio for beat/pitch detection
# apt install aubio-tools
# BPM detection
aubiotempo input.wav
# Pitch/key detection
aubiopitch -i input.wav -p yinfft
# Onset detection (transient markers)
aubioonset input.wav
# Generate stats (includes RMS, peak, DC offset)
sox input.wav -n stats 2>&1
| Format | Sample Rate | Bit Depth | Use |
|---|---|---|---|
| CD | 44.1 kHz | 16-bit | Consumer playback |
| DVD | 48 kHz | 24-bit | Video soundtrack |
| Hi-Res | 96 kHz | 24-bit | Audiophile streaming |
| Studio | 96-192 kHz | 32-bit float | Recording/mixing |
| Podcast | 44.1 kHz | 16-bit | Voice distribution |
| Phone/VoIP | 8-16 kHz | 16-bit | Voice calls |
# Downsample from 96kHz/24-bit to 44.1kHz/16-bit with dither
sox input_96_24.wav -r 44100 -b 16 output_441_16.wav dither -s
# Same with ffmpeg
ffmpeg -i input_96_24.wav -ar 44100 -sample_fmt s16 -af "dither=method=triangular" output.wav