True-tempo estimation found the tempo was never lost: the combs recover the exact declared bpm, and the same-day correction rebuilt the grid story from the audio up

The chart-draft pipeline, chartless

Study Overview

The Registered Design

P1, the chord-change joint comb, searched period across the declared value plus or minus 10 percent jointly with phase, two-stage, minimizing the circular distance of the shipped reader's grid-free segment boundaries to the lattice; P2 ran the same search on Study 22's onset objective. Arm B3 re-ran the chartless draft pipeline with P1's fitted grid, the period the only variable against 23b's blind arm. Gate one asked whether the fitted grid reaches the in-run uniform ceiling on the aligned truth; gate two asked whether the draft comes within 5 points of the grid-given baseline on every song.

The First Reading

Gate two failed as a whole and the split looked like the finding: peaceful-dreams reached 148 of 153 roots chartless, within 2 points of its grid-given 151, with 18 of 62 dialect bars written exactly, above grid-given's 11 and the best color number the program has measured; mop-heart reached 136 of 140, above its 131 baseline. Letter-lands and eat-all-the-clocks stayed collapsed, and the in-run least-squares uniform fit to the aligned truth itself hit only 3 to 17 bars per song, which the bench published as the deep fact: the tempo is not a constant. Gate one passed hollow, its least-squares ceiling disclosed the same day as the wrong objective for a hit-rate gate.

The Correction Rebuilt the Story From the Audio

Kevin's correction, that the recordings are precisely metronomic, was verified within the hour by a comb sweep over the raw onset envelope with no alignment involved: razor-sharp peaks at 131.97 to 131.99, 120.00, 144.96 to 144.98 and 223.82 to 223.97 bpm against declared 132, 120, 145 and 224. Re-scored against those true periods, this study's P2 had found the exact tempo on all four songs, 0.02 percent or better, and P1 on three of four; the bench had graded both against its own broken ruler and reported perfection as failure. The ruler's damage was then localized: its bar geometry drifts up to 22 seconds while its per-bar bins stay musically chart-aligned (so the program's chord grades stand), the 7.15 and 6.5 second 'intros' on letter-lands and peaceful-dreams were four-bar start errors (both recordings open on chart bar one, which is what actually collapsed 23b's tap arm and this study's letter grades, by whole-bar index shifts), and eat-all-the-clocks' chart needs 172 seconds of audio against a 147.5 second recording, an arrangement mismatch no grid can chart.

What Survives and What Comes Next

Surviving on the positive side: the dialect result (18 of 62 exact on a chartless grid), the onset joint comb as an exact mechanical tempo check, and the shipped reader's near chart-grade inputs under the true grid (70 of 76, 134 of 153, 111 of 140 roots on the arrangement-matched songs). Withdrawn: both tempo verdicts and every geometric use of the alignment truth; the standing lesson entered the program's ledger as 'bins only, never geometry.' The corrected grid guidance to the app: use the declared bpm, it is exact; anchor bar one near t equals zero; the open problem is chart-versus-recording arrangement. Study 25, the corrected end-to-end draft on the true grid, was registered as the follow-up with the prediction that the three arrangement-matched songs reach baseline.