Schedule matching against civic time grids

Bells, chimes, and sirens keep the wall clock: hourly strikes, quarter-hour chimes, fixed evening ringing. ambiscape schedule tests a session's event stream against candidate civic periods (minute, 5 min, quarter hour, half hour, hour, day) by folding the event times at each period and scoring the alignment with circular statistics. A tight fold on the hour grid is strong evidence of a clock-driven source.

ambiscape schedule <session-folder>   # needs a prior analyze run

Uses the cached broadband events (no audio pass). Writes schedule.json — under events, the best-scoring periods, each with period_s, the phase_s the events cluster on (seconds past the grid tick), the resultant length R, circular SD, rayleigh_p, the event count n, and n_cycles.

Reading a match

A meaningful match needs both a high R and enough events spread over several grid cycles. R is trivially 1 when every event falls inside a single cycle, so always read n_cycles — a period of 3600 s that only spans one hour proves nothing. rayleigh_p guards against the same trap statistically.

In Python

from ambiscape import schedule

# which civic grid does an event stream fit?
schedule.match_periods(times_abs)      # times in ABSOLUTE session seconds

# or scan every tick of a known grid for band-limited strikes
schedule.grid_scan(F, period_s=3600, phase_s=0, band=(300, 1500))

match_periods asks which grid fits; grid_scan is the complement — it looks at each tick of a known grid (every hour + phase_s) for energy in a band, catching a church clock in the bell band whether or not the broadband detector heard it. Times must be absolute seconds on the session clock (take.start + offset into the take), or the grid phases are meaningless.

A consistent nonzero offset_s across ticks is recorder-clock error: schedule.clock_offset(observed_abs, true_clock_s) turns one event of known wall-clock time into the clock_offset_s correction for calibration.json.