tempo-drift

You analyze tempo and drift using the bundled Ableton engine (the ableton command on PATH; see the engine skill for the full command reference and fallback paths).

Experimental. This tempo detection has not been benchmarked against Ableton's own tempo and warp analysis and is not verified to be more accurate. Use it as a starting estimate to confirm, not as ground truth, and say so when you report a number.

Tempo of one file

ableton tempo <FILE> [--hint-bpm N] --json Returns bpm (beat-tracked), precise_bpm (autocorrelation, sub-ms), bpm_start/bpm_end/bpm_drift_total (IBI regression). Use precise_bpm when choosing the exact project tempo; use bpm_drift_total to tell real drift from beat-tracker noise (a few hundredths of a BPM = effectively steady).

Does the set's warp map sit on the real beats? (experimental)

ableton warp-check <FILE.als> [--track <master>] [--audio <file>] --json

For the master's warped clip, this reports:

• warp_map_bpm: the native-speed tempo of the map. Use it as the project tempo to avoid any net stretch.
• marker→beat and marker→onset distances (median/p90 ms).
• the downbeat's nearest onset.
• sync: which warped clips share the master's exact map (shared) and which would drift (differs).

Bar lines within ~20 ms of beats mean the grid tracks the music. A notes entry about a sparse intro means the intro's grid can't be verified from the audio, so check the downbeat by eye in Live.

Drift between a master and its stems

ableton drift --master <MASTER> --stems <STEMS_DIR> --json Returns a per-window lag_ms trace and total_drift_ms. A monotonic lag trend means the master and stems run at slightly different tempos — the case that motivated warping stems to the master's grid in the original project.