Rough couple of days, in the best way. A wind-and-altitude mystery turned into a proper investigation, and it surfaced a handful of small measurement mistakes hiding in plain sight. None of them were huge on their own, but I'd rather find them myself than have someone else point them out first.
The continuous 1-second sound archive, the same one built a few days ago to let an independent researcher cross-check our numbers, ended up cross-checking us instead. Two flights this week had their sound peak logged wrong, one because a service restart cut the tracking thread short, one because the ADS-B position arrived late and got matched to the wrong moment. Both are now corrected in the data, with the original readings kept alongside the fix rather than erased. Same rule as always: a mistake found and written down openly beats one quietly smoothed over.
Vibration just got the same safety net the sound already had, an always-on archive that can be checked after the fact instead of trusting a single live reading. And the station's radio receiver, the one that turns aircraft into data points in the first place, got swapped out this evening for a spare that's been sitting ready since August, after the corrupted-reading rate crept up higher than it should.
On a lighter note: the Captain's Log itself just went multilingual. Every entry here now reads in Portuguese and French too, translated with the same care as the rest of the site, not just Google-Translate-and-forget. Forty-two backdated entries got the treatment, and every new one will from here on.
And further afield, a researcher near Milan running his own network of noise sensors reached out after seeing this project covered on Euronews. Different city, same instinct: measure it yourself, show your work. A data exchange between our two stations is in progress.
Still a few loose ends before this week counts as fully closed. But the instruments agree with each other more than they did a week ago, and that's the only scoreboard that matters here.