18 scenarios, built from the same set of official acoustic nodes.
Automotive · appliances · rotating machinery · environmental noise.
Resample the time-frequency spectrum into the speed-order domain to see at a glance which order sounds at which RPM.
Turn the inspection of hundreds or thousands of motors/pumps/fans into a single chart: compute velocity RMS, apply Zone thresholds, and output a red-green status.
Envelope demodulation surfaces early bearing faults buried in broadband vibration, matching BPFO/BPFI/BSF/FTF characteristic frequencies on the envelope spectrum.
Multi-mic synchronous capture + A-weighting + one-third octave, synthesized into a single sound-power level L_WA with a PASS/FAIL result.
The Loudness / Sharpness / Roughness / Tonality quartet quantifies "premium vs. cheap" perception into repeatable objective metrics.
A/C/Z weighting, Fast/Slow/Impulse, Leq/Lmax/Lpeak/statistical levels, and day-evening-night grouping are baked into the nodes for consistent re-measurement.
Split road-noise and wind-noise segments by GPS speed, then build a "speed × band" heatmap with consistent conventions to fairly compare states A and B.
8h–24h outdoor recordings: equivalent level + narrowband-hum prominence (Tonality/TNR) + day-evening-night evaluation, built once and rerun long-term.
E-drive whine · AVAS compliance · the quiet sounds inside a pure-electric cabin.
Electromagnetic orders, mesh orders, and PWM sidebands are intermixed; extract orders separately in the mechanical and electrical domains and overlay them in a two-color Campbell diagram.
Segment by speed, auto-fit the fundamental-frequency drift slope, check sound-level limits, and produce a PASS/FAIL result with a type-approval certificate page.
With the engine's masking gone in a pure EV, door-closing, air-vent, and rattle sounds are amplified; first auto-capture transient events, then run psychoacoustic evaluation.
Run a short sweep on each unit at end of line, auto-extract whine orders and tonal prominence by RPM segment, and return PASS/FAIL within cycle time.
For the creaks and clacks (BSR) on a bumpy road or shaker table, auto-capture transients + psychoacoustic characterization to turn subjective complaints into comparable objective items.
The same operators, carried into healthcare · ecology · power · security.
Carry the "time series + spectrum + anomaly" operators over to physiological signals: S1/S2 segmentation + murmur spectrum + AutoML normal/abnormal triage.
Unattended long recordings: Mel spectrum + structure-tensor texture + AutoML multi-class species identification, producing an activity-rhythm data foundation.
Treat the "current WAV" as a time-series signal: FFT extracts harmonic spectral lines, computes each-order amplitude and THD%, and checks against the standard limit table.
For 20–40 kHz ultrasonic leaks inaudible to the human ear: baseline statistics model the normal distribution, and Z-Score auto-flags anomalies and alerts.
The harmonics and modulation of a core's 100/120 Hz hum reflect loosening and DC bias magnetization: modulation spectrum + baseline deviation for 7×24 online early warning.
Nodes are general-purpose operators for time series, spectra, anomalies, and compliance.
Just recombine them for your data and standards. Tell us what you need and we'll build it together.