Drop a file or capture from the microphone to begin.
—
A measurement bench for any recording you drop in or capture. It reports peak and RMS levels, DC offset, noise floor and signal-to-noise, clipping runs, stereo correlation, spectral centroid and rolloff, the upper band edge that reveals lossy encoding, and mains hum resolved to 50 or 60 Hz.
Analysis runs in a worker over the decoded buffer, so the same file measured twice gives identical numbers. Every finding is shown with the measurement it rests on, and a selection can be re-measured on its own. It describes recordings; it does not identify their causes.
The level of the quietest sustained sound in a recording — hiss, room tone, electrical noise. Signal-to-noise ratio compares it with the loudest material. A sound that barely rises above the noise floor is very hard to interpret reliably.
A tone at the frequency of the electricity supply — 50 Hz in most of the world, 60 Hz in North America and parts of South America and Asia — and its harmonics. It leaks into recordings from cables, chargers and nearby equipment, and the analyzer reports which of the two it found.
It looks for runs of consecutive samples at or near full scale. Clipping flattens the top of the waveform, adding harsh distortion that can make ordinary sounds seem strange.
No. Decoding and analysis run in your browser, in a background worker. The same file measured twice gives identical results.
No. It describes recordings — level, frequency content, distortion, encoding — and shows the measurement behind each finding. Identifying a source needs context the audio alone does not contain.
This is an instrument and a sound tool, provided for entertainment and experiment. It describes what it measures and does not identify the cause of anything it shows.