Compare two microphones by numbers, not impressions: mic-compare
Record the same phrase on two mics at matched gain and let the agent report response, noise floor, per-band SNR, plosives and sensitivity, and what EQ can still fix.
Use the mic-compare skill on these two takes: {mic_a.wav} and {mic_b.wav}. First check the recording is valid: matched preamp gain, peaks between -12 and -6 dBFS, no clipped samples. Then report LUFS, SNR, spectrum shape normalised to 300-3000 Hz, voice minus noise per band, hum, and plosives per minute. End with a verdict: what EQ can fix, what is baked into the mic.
In Claude Code, paste the prompt as a message.
Install
npx skills add fortunto2/superduper-dsp --skill mic-compare# or the whole set as a Claude Code plugin (commands then look like /superduper-dsp:mic-compare):
claude plugin marketplace add https://github.com/fortunto2/superduper-dsp
claude plugin install superduper-dsp@superduper-dsp --scope userSteps
- Record both mics saying the same 30-plus second phrase at the same distance and the same preamp gain, with real pauses (the pauses measure the noise floor).
- Install the skill. The measuring code runs with uv: numpy, scipy, soundfile and pyloudnorm.
- Paste the prompt. Read the validity check before the verdict.
Tips
- Matched gain decides everything. In the author's test a mic looked 5.7 dB cleaner with free-run gain and 2.8 dB noisier once gain was matched.
- A boost raises signal and noise equally: an EQ shelf cannot fix a worse per-band SNR.
- Room pickup at 300 to 2000 Hz cannot be EQed out; it sits under the voice.
| works with | Claude Code: Runs as a skill or a prompt in the terminal; can read your files and write the result. |
| cost | Free. No API, no key. |
| status | Unverified: written from the technique, not yet run end to end by us. |
| source | mic-compare skill in superduper-dsp (MIT) |
| license | CC BY 4.0 for the text here; skills are MIT. |
| updated | 2026-11-10 |
| for agents | mic-compare-two-microphones-by-numbers.md |
Related
Voice cleanup checklist: room tone, de-ess, noise gate, in that order
Six passes that turn a phone or lav recording into a usable voice track, each with the ffmpeg filter and a starting parameter to adjust by ear, in the order that keeps the gate off word endings.
Mix and master by measurement: sdsp-mix
An agent mixes the way an engineer would: render, measure against a commercial reference, change one thing. Mix map of who masks whom, kick and bass ducking, mastering to a loudness target.
Nine agent skills for mixing, mastering and making music (superduper-dsp)
A Claude Code plugin that teaches an agent to mix by measurement, build songs in REAPER, render plugin chains headless, turn audio into MIDI and tear down a reference track.
Recording to MIDI: pick the model by material, then check it
Three local transcription models fail in different directions. The skill picks one by what the material is, runs it, and checks the result before you trust it.
Render plugin chains from a text file, no DAW: sdsp-chain
One TOML file describes tracks, effects, sidechains, automation and playable instruments; a Rust command renders it to WAV and prints LUFS and true peak per stage.
Mashups and ciphers with a beat grid: sdsp-mash
Pick pairs by BPM and key, split stems, stretch only the beat, and render one beat with verses from other songs, or a beat switch per verse, from a TOML file.