Signal-to-Sound Lab
Turn two normalized time series into a visual comparison and a stereo musical sketch. This is an offline educational instrument, not an OBD decoder, calibrated oscilloscope, diagnostic decision system, or medical device. Sound is created locally after you press Play.
1 · Choose evidence
Required columns: time_ms,signal_a,signal_b. Signals must be normalized 0–1, timestamps strictly increasing, and duration ≤ 60 seconds. Maximum 2,500 rows and 128 KiB. Imported data is never labeled measured or authenticated.
Demo is deterministic and clearly labeled synthetic.
2 · Inspect both signals
3 · Hear the comparison
Ready. Audio starts only after a click.
A maps 0–1 to 146.8–440 Hz (left); B maps 0–1 to 293.7–880 Hz (right). Loudness is bounded; signal amplitude changes pitch, not volume. The app uses at most 128 time bins. Silence does not imply zero voltage or sensor failure.
Verification and limitations
This display preserves relative ordering and window averages, not high-frequency spikes. Downsampling may hide brief transients. Imported CSV can be simulated, edited, or mislabeled; provenance is not authenticated. No file is transmitted, no remote scripts are loaded, and no measurement is acquired from a vehicle. Use a calibrated scope and manufacturer procedures for actual diagnosis.
CSV example: time_ms,signal_a,signal_b 0,0.10,0.90 100,0.30,0.70 200,0.50,0.50
Reproducibility: demo uses fixed mathematical functions. Error cases fail closed. Export writes the current normalized samples only. For a real workflow, preserve the original capture, sampling rate, units, sensor calibration, timestamp origin, and checksum separately.