Datasets:
audio audio |
|---|
nightjar flight — 2026-08-04, fourth sortie
Last sortie of a four-flight day, and the one with the most time under an optical track. Same rig and site as the other 2026-08-04 sessions. GPS truth: 3,572 valid UTC stamps of 3,573.
| airborne | 357 s |
| max height | 26 m |
| max slant range | 244 m |
| median elevation angle | 5° |
| path flown | 894 m |
The interesting number
53 cue candidates, and 43 of them were evaluated at the 14.6 dB threshold. The deployed presence gate is state-dependent — strict (14.6 dB) while nightjar reports TRACK, relaxed (10.0 dB) while it is searching — so a candidate stamped 14.6 means the tracker was locked at that moment. Only 10 candidates saw the relaxed threshold. Compared with the third sortie, where all 8 candidates were at 10.0 dB and the tracker was never locked when a cue was evaluated, this session spent substantially more of its life holding a track.
That is a suggestive contrast, not a controlled one: the two sorties differ in duration (357 s vs 200 s), range (244 m vs 142 m) and path length (894 m vs 357 m) as well.
audio/cue_log.jsonl stamps thr_db and nj_state on every candidate, so each cue is
attributable to the threshold in force. 4 chirp-rejects. Two start records appear because
the cue daemon was restarted into the session after a device-contention race at launch (the
daemon is the UMA recorder — the array is single-open).
Caveat that applies to the whole day
⚠ 26 m altitude at 244 m range is a ~5° elevation angle — the drone was on the treeline, against foliage rather than sky. Three of the day's four sorties flew this profile. It is a hard background for the detector and it sits at the null of a horizontal array's elevation response (sensitivity goes as sin(el), so ~0 at the horizon). The 50–150 m detection question that motivated the day's work therefore remains unexercised.
Contents / gotchas
audio/uma16_<epoch>.wav UMA-16, 16 ch, S16_LE audio/sb_<epoch>.wav SB-POLARIS, 8 ch, S32_LE
audio/cue_log.jsonl cue gate log (thr_db, nj_state per candidate)
video_segments.tar 1080p native + 768x432 substream, 60 s MPEG-TS
blackbox/state.jsonl tracker FSM ~2.9 Hz blackbox/frames.tar 2,856 stills f<epoch_ms>.jpg
dji/*.csv, *.txt decoded + original FlightRecord (CUSTOM.dateTime = true UTC)
UMA-16 0-based channel 9 is a dead electronic-floor channel — exclude it, run on 15. SB-POLARIS has 3 live capsules of 8; detect them by RMS (live ~−70 dBFS, dead ~−190) rather than hardcoding. ⚠ The arrays use different sample widths — UMA S16_LE, SB S32_LE; reading both as int16 scrambles SB channel identity and destroys its spectrum.
Calibration as flown: az_offset_deg 301.37, invert_az true. Cue producer: cheap daemon
(SUBHUNT=0). Site notch 100–200 Hz. Start pose pan 178.9 / tilt 45, pointed at the JBL
beacon. Cue-gated blob seeding was present in the build but disabled for this flight.
GPS-referenced pixel work
Placing the drone in the image from GPS reaches a floor of about 5° of bearing error on this rig's data — the DJI's own ±2–5 m of position is 4–9° of bearing at the 20–45 m ranges where the optical correspondences live. Treat any in-frame or pixel-level claim derived from GPS here as approximate.
- Downloads last month
- 67