메뉴
FFT-based sanity check for audio output files — verifies fundamental frequency, THD, SNR, clipping, and silence. Use after offline C tests or WASM OfflineAudioContext captures to confirm a signal chain is producing correct output.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Scan Claude Code JSONL session transcript files for behavioral patterns — loops, errors, expensive turns, trial-and-error signatures. Returns structured findings for the improvement-analyst agent.
Run a Criterion benchmark on a Rust DSP crate and assert against a checked-in baseline. Establishes whether the implementation fits the real-time budget and catches performance regressions — the Rust equivalent of `perf stat` against the audio buffer time.
Regenerate a C header from a Rust crate via cbindgen and diff it against the committed header. Fails if the generated output differs — catches "Rust signature changed but the C header was not regenerated" before it reaches the C side.
Run `cargo +nightly miri test` against a Rust DSP crate to detect undefined behaviour the compiler missed — out-of-bounds access, use-after-free, uninitialised memory reads, invalid pointer arithmetic. Miri is the cheapest sound UB detector for Rust; this skill wraps it with the right isolation flags for audio DSP crates.
Audit a Rust DSP crate for real-time-safety violations — verifies `#![no_std]`, no `alloc` dependency, `panic = "abort"`, runs targeted clippy lints, greps for banned constructs in the hot path, and scans the compiled LLVM-IR for allocator symbols. Designed to be run on any Rust DSP crate before merge.
Opinionated wasm-pack wrapper that builds a Rust crate to wasm32-unknown-unknown with project-pinned flags (SIMD, LTO, panic=abort), applies a wasm-opt post-pass, and auto-invokes wasm-module-inspect on the result. Single entry point so the agent never reaches for raw wasm-pack and forgets a flag.
| name | audio-fft-sanity |
| description | FFT-based sanity check for audio output files — verifies fundamental frequency, THD, SNR, clipping, and silence. Use after offline C tests or WASM OfflineAudioContext captures to confirm a signal chain is producing correct output. |
| allowed-tools | Bash |
| argument-hint | <audio-file.wav|float32> [--freq HZ] [--sr HZ] [--thd-max PCT] [--snr-min DB] [--channels N] |
Run a suite of FFT-based assertions on an audio output file to confirm a signal chain is behaving correctly. Suitable for both offline C test output and WASM OfflineAudioContext captures.
numpy and scipy: pip install numpy scipyArguments are in $ARGUMENTS. Pass them directly to the analysis script:
python3 "$(dirname "$0")/analyse.py" $ARGUMENTS
| Check | What it detects |
|---|---|
CLIPPING | Any sample magnitude > 0.99 — indicates gain staging problem or overflow |
SILENCE | RMS below -60 dBFS — catches zero-output bugs (uninitialised buffer, wrong pointer) |
FUNDAMENTAL | Detected peak frequency vs --freq (if given) — verifies oscillator/filter tuning |
THD | Total harmonic distortion (harmonics 2–5 vs fundamental) — catches nonlinear distortion or aliasing |
SNR | Signal-to-noise ratio estimate — catches noise floor issues or quantisation problems |
0 — all checks passed1 — one or more checks failed# Basic smoke test — check for clipping, silence, SNR, THD with defaults
python3 analyse.py output.wav
# Verify a 440 Hz sine from a synth, strict THD limit
python3 analyse.py output.wav --freq 440 --thd-max 1.0 --snr-min 60
# Raw float32 file at 48 kHz, stereo interleaved (uses first channel)
python3 analyse.py output.f32 --sr 48000 --channels 2 --freq 1000
Synthesizer test: Run the synth at a known pitch (e.g. MIDI A4 = 440 Hz), render N frames to a float32 file, then:
python3 analyse.py synth_output.wav --freq 440 --thd-max 2.0 --snr-min 50
Effect processor test (e.g. filter, EQ): Feed a broadband noise or sweep, capture output — omit --freq and check only for clipping and SNR:
python3 analyse.py filtered_output.wav --snr-min 30
Impulse response capture: Feed a single-sample impulse, check output is non-silent and not clipping:
python3 analyse.py ir_output.wav