Kiev, Aug 2011 vs Conklin, Alberta, Jan 2012
Two "strange sounds in the sky" uploads, five months and 8,000 km apart. Use the controls to test whether they're the same recording.
Waveform match (whole clip)
-
Correlation between the two waveforms, sample by sample, from 0 (unrelated) to 1 (identical). Kiev as uploaded scores 0.25 against Conklin; tone-matched, 0.70, and 0.8–0.9 during the loud bursts. At any wrong offset it stays under 0.02 (tested at 160 offsets). Slide the offset away from 0 and watch it collapse. The quiet stretch at the start (about 5–15 s) scores low on its own because there's little of the burst sound in it.
■ Kiev (as currently matched), live spectrogram, 0–2.5 kHz
■ Conklin, live spectrogram
Loudness of each burst over the 45 s (Kiev / Conklin), playhead in white
What you're hearing. Conklin's soundtrack starts 15.6 s into the Kiev upload. Its bass is cut by about 25 dB, which is why the two sound different at first. "Tone match" applies the same filter to Kiev, measured from the two recordings themselves. In Switcher mode the audio flips between the two every few seconds; a different recording would jump at every switch. Split puts one in each ear (headphones).