docs(ear-feedback): superfreak/hammer kit audition — sub-only bass, BCDA loop…

docs(ear-feedback): superfreak/hammer kit audition — sub-only bass, BCDA loop rotation, demucs vocal high-pass

PLN's on-disk verify of the forged kits, with measured confirmation of all
three reactions: bass_deep is a pure-sub layer (99% <150 Hz at −12.8 dBFS —
equal-loudness invisibility, not a level bug); loops are bar-exact but
musically ROTATED (strongest onset lands on beat-slot 1/2/4 instead of 0;
other_skank opens −23 dB below its own peak 30 ms — the 'cut DABC' feel);
vox thinness is htdemucs routing vocal low end into bass/other (vocal stems
carry ~0 energy <150 Hz). Fix directions logged: rotate-to-downbeat export
post-pass, BS-Roformer vocals re-cut, harmonic-saturation bass variant.

armada/tide-table/performance_notes.md

@@ -171,3 +171,29 @@ Lessons:
 **Open:** the real 38C3 PunkAChien ("Pitbul Punk") is still unfound — Take35 eliminated.
 Candidates left: Take36 / the 61:41 "House of Tea" set, Take37/38 (Chaos Music Club),
 or it was never recorded as a standalone. Hunt deferred behind the bleed-detector build.
+
+**Foundry kit audition — superfreak/hammer (2026-07-11, PLN on-disk verify).** Ear-report
+on the forged kits: (1) "lots don't loop well or are not audible — bass_deep is not
+audible"; (2) "other_skank is good but seems not starting right — loops feel 'timed ok'
+but cut BCDA or DABC, not at a good start"; (3) "vox superfreak/kinkygirl are very
+high-pass filtered? some bass split? is this demucs?"
+
+Measured confirmation (all three are real, none is a duration bug):
+- **bass_deep inaudible** = pure-sub sample: RMS −12.8 dBFS (loud!) but 99% of energy
+  <150 Hz (75% in 60–150), centroid 478 Hz. Physically invisible on small monitors —
+  equal-loudness, not level. It's a sub LAYER (two-layer bass idiom), not a standalone.
+- **BCDA/DABC rotation confirmed**: strongest onset beat-slot inside each loop —
+  drums_dub slot 1/16, bass_deep slot 4/16, other_skank slot 2/8; other_skank's first
+  30 ms is −23 dB below its loudest 30 ms (starts mid-phrase). Root cause: the PLP beat
+  grid has no downbeat anchor, candidates start on ANY beat. The loops are bar-exact
+  (tempo-locked) but musically rotated.
+- **vox high-pass = demucs htdemucs**, not our cuts: the vocals STEMS have ~0 energy
+  <150 Hz (superfreak 0.1% in 60–150, hammer 0.9%) — vocal chest/low end is routed into
+  the bass/other stems by the model. superfreak bass stem holds 70% of its energy at
+  60–150. Lyric cuts inherit the thinness.
+
+Fix directions (foundry): rotate-to-downbeat export post-pass (a clean-seam bar-exact
+loop can be rotated freely — the new seam is an interior continuity point → zero risk,
+pick rotation with strongest attack at 0); vocals via engine2 BS-Roformer (keeps low
+end far better than htdemucs) then re-cut the same phrase bounds; bass_deep → optional
+harmonic-saturation "audible" variant, or accept as sub layer + in-session shape.