“Asleep in the Trunk” · the drum stem split into 5 individual kit pieces · solo, mute, volume per piece.
The Moises drum stem was separated into kick, snare, toms, hi-hat, cymbals by the per-lane DrumSep pass (MDX23C). Every piece below was measured for real drum signal before it went live — band profile and loudness, not file count. Solo a piece to isolate it; the five pieces together reconstruct the kit. The gold Full Kit lane is the original stem, muted by default — un-mute or solo it to A/B against the split.
Each separated piece measured over the full 4:39 with a windowed-FFT band profile (no resampling). The gate is real drum signal in the expected band, not a silent or garbage stem. All five carry genuine signal — none hit the −90 dB “empty toms” failure the separator can produce.
| Lane | RMS | Peak | Active | Dominant band (energy) | Verdict |
|---|---|---|---|---|---|
| Kick | −18.4 dB | −1.5 dB | 82% | sub 20–60 Hz (62%) + low 60–120 (29%) | PASS |
| Snare | −24.5 dB | −0.4 dB | 76% | low-mid 120–400 Hz (86%) | REAL — body-dominant* |
| Toms | −23.8 dB | −2.1 dB | 35% | low 60–120 Hz (57%) + low-mid (16%) | PASS |
| Hi-Hat | −42.1 dB | −15.1 dB | 53% | high 5–20 kHz (81%) | PASS |
| Cymbals | −36.4 dB | −8.9 dB | 63% | high 5–20 kHz (55%) + high-mid (44%) | PASS |
| Full Kit (ref) | −16.3 dB | −0.25 dB | 99% | sub 20–60 (43%) + low (29%) + low-mid (23%) | parent stem |
*Snare captured the shell body (150–250 Hz) but is thin on the wire crack above 2 kHz — real drum signal, worth a tighter re-pass if the snap matters. All lanes: 44.1 kHz, 4:39, identical length to the parent (duration test passes).