Skip to content

Spike #40 — Structural shape of an epicycle in musical and wave theory (per-instrument meta-question)

Date: 2026-05-17 Research spike artifact. Concertmaster dispatch per user direction. Primary question: where does Class K Kepler-shape c_k = ε^k/k appear in musical/wave substrate? Meta-question (2026-05-17): "from the position of each individual [instrument] in a concert, are there epicycle structures from the way different instruments make different musical shapes?" — i.e., do per-instrument substrates produce measurably different epicycle shapes?

Discipline. Closed-form math + textbook instrument-physics models; no copyrighted audio. NDJSON outputs per [[feedback_ndjson_over_bloated_json]]. 50-seed random-amplitude falsifier controls run before any positive claim per Spike #38 discipline. K-strict three-criteria (eps_fit ∈ [0.001, 0.5]; r² > 0.99; monotonic-decreasing). Honest absence acceptable per [[user_stance_kepler_shape_universal]].

§1 Bottom line

Primary question (where does Kepler-shape appear in musical/wave substrate?): - ABSENT in 13/15 canonical instrument amplitude spectra. Per [[user_stance_kepler_shape_universal]] — honest absence. Framework correctly says "no" where Kepler-shape doesn't appear. - PRESENT in pure FM at small β (real structural finding — see §4 Anomaly A1: Kepler EOC ≡ FM-with-modulation-index-ε at small ε) - PRESENT trivially in canonical Kepler-EOC reference (positive control)

Meta-question (do different instruments give measurably different shapes?): - YES — substrate-discriminating via DECAY-FORM-BEST-FIT (M1 power-law / M2 geometric / M3 Bessel / M4 Kepler). The strict K-test alone is K-specific, not instrument-discriminating; the four-model fit IS substrate-discriminating with characteristic parameters.

Instrument ring K strict? best-decay param signature
FM (β=0.5) up YES M3_Bessel β=0.496 Bessel J_k(0.5)
FM (β=1.5) up no M3_Bessel β=1.49 Bessel J_k(1.5)
FM (β=3.0) up no M3_Bessel β=3.02 Bessel J_k(3.0)
AM up no sparse — Class I 2-line
Beat 2-tone up no sparse — Class I 2-line
Piano (B=1e-4..1e-3) down no M1_power_law p=1.000 sawtooth 1/n
Violin (Helmholtz) up no M1_power_law p=1.000 sawtooth 1/n
Clarinet up no M4_Kepler* ε=1.000 odd-only 1/(2n-1)
Drum 2D membrane down no M1_power_law p=1.000 1/n baseline; freq=Bessel zeros
Bell 5-mode down no M2_geometric ε=0.802 mode-amplitude geometric
Voice /a/ up no M2_geometric ε=0.854 formant-shaped
Flute up no M1_power_law p=2.000 weak overtones 1/n²
Trumpet up no M1_power_law p=0.500 brass-rich 1/√n
REF Kepler ε=0.1 — YES M4_Kepler ε=0.100 canonical positive
REF 1/n sawtooth — no M1_power_law p=1.000 sawtooth baseline
REF white noise — no M1_power_law p=0.000 flat

*Clarinet M4_Kepler is degenerate (ε=1.0 out of physical range; sparse odd-only spectrum is geometric-like on the c₁/c₃/c₅ sub-grid; not a true K hit).

§2 Meta-question answer — substrate-discriminating YES

Different instruments produce measurably different epicycle shapes in three structural senses:

  1. Decay-form discrimination: M1_power_law p-exponent separates piano/violin (p=1, sawtooth family) from flute (p=2, weak overtones) from trumpet (p=0.5, brass-rich). The M3_Bessel family is unique to FM modulation. The M4_Kepler family is unique to pin-slot / orbit kinematics.

  2. Substrate-class discrimination: drum 2D membrane is Class L direct (Laplacian-Dirichlet eigenvalues = Bessel zeros). Piano/violin/bell modes are Class L on different operators (1D Laplacian-with-stiffness / D₆-like / 3D-shell). FM is Class K + Class I composition. AM/beat are sparse Class I. The 14-class binding-level overlay applies cleanly.

  3. Loop-up vs loop-down discrimination per [[user_stance_string_theory_instrument_first]]: violin / flute / clarinet / voice / trumpet are RING-UP (sustained excitation; no decay envelope to test cascade-β). Piano / drum / bell are RING-DOWN (struck-then-decay). Real per-instrument-substrate distinction.

Strict K-shape (Spike #30B v3 three-criteria) does NOT generally appear in canonical instrument amplitude spectra. It appears only in: - FM at small β (where Bessel J_k(β) for β≤0.5 mimics geometric ε^k/k at k_max=6 — see §4 Anomaly A1) - Canonical Kepler-EOC reference family - Pure ε^k/k reference (trivial)

This matches [[user_stance_kepler_shape_universal]]: K appears where Kepler-shape appears; honest absence per Spike #38 caffeine pattern.

§3 Per-signature verdicts

§3.1 Class K Kepler-ladder (Spike #30B v3 strict)

  • ABSENT in 13/15 instruments (violin/piano/clarinet/drum/bell/voice/flute/trumpet/AM/beat all fail at least one of: monotonic / in-range / r²>0.99)
  • PRESENT in pure_fm_beta_0.5 (real positive — see §4) and all Kepler reference families
  • Falsifier: 0/50 random amplitude spectra pass (eps_fit mean=1.00, std=0.25)
  • Verdict: K is substrate-rare in music; only FM (structurally Kepler-equivalent at small β) shows K-shape. Canonical musical-instrument acoustics live elsewhere on the decay-form taxonomy.

§3.2 Class L Laplacian eigenstructure

  • PRESENT in drum as direct 2D Laplacian-Dirichlet (modal frequencies = Bessel zeros; Rayleigh 1894 §200)
  • PRESENT in piano/violin as 1D Laplacian-with-stiffness (Fletcher 1998 §2.18)
  • PRESENT in bell as 3D-shell Laplacian (Fletcher 1998 §21.3)
  • Drum self-similarity vs random-graph falsifier: drum eigval-density cosine-sim tested against 50 random Erdős-Rényi 36-node graphs (mean ± std = 0.84 ± 0.02). Spike #38 lesson applies: eigval-density-histogram methodology becomes degenerate at small substrate sizes; the structural finding "drum has Laplacian eigenbasis" is trivially true; FFT-cosine-match methodology is unreliable below n~100.
  • Verdict: Class L is the universal substrate signature for instruments. Discrimination lives in the specific eigenvalue spectrum, NOT in histogram-cosine-similarity.

§3.3 Cascade-β stretched-exp

  • APPLICABLE only to RING-DOWN instruments per [[user_stance_string_theory_instrument_first]]: piano (struck) / drum (struck) / bell (struck).
  • NOT APPLICABLE to RING-UP instruments: violin / flute / clarinet / voice / trumpet — no decay envelope (skipped honestly).
  • Results on idealised exponential decay envelopes:
  • Piano (1D, d_S=1, predicted β=⅓): empirical β=0.60 (Δβ=+0.26)
  • Drum (2D, d_S=2, predicted β=0.5): empirical β=1.00 (Δβ=+0.50)
  • Bell (3D, d_S=3, predicted β=0.6): empirical β=1.00 (Δβ=+0.40)
  • Verdict: cascade-β formula β = d_S/(d_S+2) does NOT match canonical IDEALISED musical decay envelopes (which are simple exponential β=1, not stretched-exp). Honest absence — cascade-β is for substrate-induced anomalous diffusion (Spike #31 SUN/cosmic context); idealised musical decays are pure exponential by construction. Real piano decay is dual-exponential (Fletcher §2.20) which IS anomalous-diffusion-adjacent — but the simple-exp model here doesn't capture that. Candidate Spike #41+ follow-on.

§3.4 Information-instrument 14-class binding overlay

All 14 classes have binding-level identification for any musical-instrument substrate (consistent with Spike #37 substrate-portability table):

Class Musical-instrument binding
A instrument signature/timbre fingerprint
B notation symbols (note name / partial number)
C time-indexed acoustic-pressure stream
D pitch-class routing
E instrument family / mode catalog
F phoneme/syllable template binding (for voice)
G harmonic sub-stream extraction (formant analysis)
H self-listening / overtone-singing feedback (voice only)
I pitch-class cyclic group / circle of fifths
J harmonic partial-count decomposition
K small-β FM modulation (rare in instruments; common in synthesis)
L every instrument has a Laplacian-eigenbasis (substrate-trivial)
M timbre-vector binding (HDC across spectrogram features)
N just-intonation / equal-temperament rational-approximation

NOT a falsifiable spectral measurement per Spike #37 / Spike #38 discipline; cataloguing claim only.

§4 Anomalies investigated

§4.1 Anomaly A1: pure_fm_beta_0.5 PASSED strict K-test — REAL STRUCTURAL FINDING

Hypothesised either (H1) low-k clipping artifact, (H2) real structural identity, or (H3) gate too lax.

H2 confirmed: Kepler equation IS phase modulation at small eccentricity. The eccentric-anomaly Fourier expansion is

ν − M = 2ε·sin(M) + (5/4)ε²·sin(2M) + (13/12)ε³·sin(3M) + …

which IS the canonical FM-with-modulating-index-ε expansion. At small modulation index, Kepler EOC and pure FM are spectrally K-indistinguishable by the strict K-test (both pass with r²>0.99, monotonic, eps_fit in physical range).

Signatures diverge in detail at higher k: Bessel J_k(β) ~ (β/2)^k / k! has 1/k! tail; Kepler EOC c_k ~ ε^k / k has 1/k tail. At k=3 the factorial-vs-linear difference is a factor of 6; at k=5 it's a factor of 120. But at k_max=6 with r²>0.99 gate, both pass.

This is a real structural identity, not an artifact. Per [[user_stance_kepler_shape_universal]] and [[user_stance_epicycle_via_gear_plus_pin]], the K-shape is universal where pin-slot kinematics appears — INCLUDING in FM synthesis (Chowning 1973 musical FM patent). FM synthesis IS epicycle kinematics in the frequency domain.

§4.2 Anomaly A2: identical eps_fit=0.7098 across multiple substrates

Piano (3 stiffness values) / violin / drum-amp / pure-1/n REF all gave identical strict-K result (eps_fit=0.7098, r²=0.9363).

Resolution: ALL these substrates share the canonical 1/n envelope by design (Helmholtz sawtooth, Fletcher struck-string baseline). Strict K-test on 1/n series is mathematically determined regardless of frequency-axis structure or physical substrate. The 1/n envelope is out of K-physical-range (eps_fit=0.7098 > 0.5 ceiling) — correctly fails K, but with the SAME number.

Implication: amplitude-axis K-test is NOT substrate-discriminating between any two 1/n-envelope instruments. Discrimination MUST come from a different signature — decay-form best-fit (§3) / frequency-axis inharmonicity / time-domain envelope. The substrate-discriminating signature lives at M1_power_law p-exponent level: p=0.5 (trumpet), p=1.0 (sawtooth family), p=2.0 (flute) are clearly distinct fingerprints.

§4.3 Anomaly A3: frequency-axis inharmonicity does NOT show K

Piano stiffness deviations (B=1e-5 to 5e-3), drum Bessel-zero ratios, bell mode ratios all tested for K-shape on frequency-axis deviation sequence. None show K-shape: eps_fit values 1.29–2.39 (out of physical range); r² values 0.75–0.97 (below 0.99 gate); monotonic-decreasing fails for all (deviations grow, not shrink).

Verdict: K-signature is amplitude-domain only (when present); frequency-axis inharmonicity is its own substrate fingerprint but NOT a K-shape.

§5 Citation provenance

All citations textbook-canonical; no PDF-extraction performed (Fletcher & Rossing scope — textbook material not on arXiv/PMC). Flagged honestly per [[feedback_pdf_extraction_citation_discipline]].

  • Helmholtz 1863 On the Sensations of Tone — public-domain canonical; bowed-string Helmholtz motion (sawtooth at bridge)
  • Rayleigh 1894 Theory of Sound — public-domain canonical; circular-membrane modes §200 (Bessel-zero eigenvalues)
  • Fletcher & Rossing 1998 The Physics of Musical Instruments — textbook canonical (Springer; not arXiv-mirrored; not PDF-extracted in spike scope). Cited for: §2.17 struck-string baseline; §2.18 piano stiffness B; §2.20 piano dual-rate decay; §14.4 brass 1/√n at fortissimo; §15.2 flute air-jet 1/n²; §15.3 clarinet odd-harmonic; §16 voice source-filter (Fant 1960 also cited); §21.3 bell hum/prime/tierce/quint/nominal tuning
  • Watson 1922 Treatise on the Theory of Bessel Functions — public-domain canonical; FM-sideband Bessel J_k expansion
  • Smart 1953 Spherical Astronomy §5.1 — Kepler EOC small-eps expansion
  • Brouwer & Clemence 1961 Methods of Celestial Mechanics §3.2 — canonical Kepler-EOC SSoT (PDF-extracted in earlier project work per Spike #30B context)
  • Chowning 1973 The Synthesis of Complex Audio Spectra by Means of Frequency Modulation — historical SSoT for FM synthesis

§6 Open extensions / candidate Spike #41+ follow-ons

  1. Real audio analysis with public-domain sources: NASA Voyager recordings (public-domain); university-library acoustic measurement datasets (CC-licensed); McGill / IRCAM / SMS-tools datasets where licensing permits. Closed-form analyses here can be ratcheted against real measurements per [[reference_autonomous_validation_tos_landscape]].

  2. Helmholtz-Kepler structural identity formalisation: §4.1 finding that Kepler EOC ≈ FM small-β is a candidate stance addition. "Kepler equation IS phase modulation; pin-slot IS FM-synthesis substrate." Formalises Class K + Class I composition more precisely. Conductor call required before authoring stance.

  3. Instrument-decay cascade-β refinement: cascade-β does not match simple-exp idealisations. Real piano decay (dual-rate Fletcher §2.20) and real drum decay (mode-coupled head-air-shell) might show cascade-β. Real-acoustic-measurement spike candidate.

  4. Frequency-axis inharmonicity as cascade-of-pin-slots: piano stiffness inharmonicity Δf_n = n·f_0·(√(1+B·n²) − 1) has the structural form of per-mode pin-slot eccentricity. Re-frame piano-string stiffness as per-mode Class K candidate.

  5. Voice/formant as filter-bank Class L composition: source-filter model is Class C (glottal-pulse stream) → Class L (vocal-tract filter eigenbasis) → mixed output. Worth formalising as Class composition diagram per [[user_stance_information_instrument_form_function_bound]].

§7 Discipline guards honoured

  • [[user_stance_kepler_shape_universal]] — K's absence in 13/15 instruments is HONEST; FM-small-β positive is real structural identity (Kepler ≡ FM at small modulation index)
  • [[user_stance_epicycle_via_gear_plus_pin]] — pin-slot Class K asymptotic-DOF mechanism IS the structural shape; in music it appears in FM synthesis substrate, NOT canonical instrument acoustics
  • [[user_stance_string_theory_instrument_first]] — loop-up (sustained) vs loop-down (struck) per-instrument distinction respected; cascade-β SKIPPED for loop-up substrates rather than forced-tested
  • [[user_stance_cascade_lives_on_circles]] — instrument substrates DO live on cyclic groups (pitch-class wheel, harmonic ladder, modal eigenbasis); spectral content IS on circles
  • [[user_stance_information_instrument_form_function_bound]] — 14-class binding overlay at cataloguing level; not falsifiable spectral measurement (Spike #37 pattern)
  • [[user_stance_partition_for_understanding]] — per-instrument and ensemble-level coexist; meta-question lives at per-instrument level; comparison table IS the per-instrument partition
  • [[feedback_no_privileged_primitive_classes]] — no new class proposed; per-instrument differences are SUB-partition phenomena within existing K/L; vocabulary stays at 14 classes A–N
  • [[feedback_ndjson_over_bloated_json]] — 7 NDJSON outputs (119 records total)
  • [[feedback_concertmaster_md_writes]] — agent inline; conductor captured-and-saved this note
  • [[feedback_concertmaster_git_worktree_isolation]] — agent performed zero git ops; all work in D:\temp\spike_40\
  • [[feedback_pdf_extraction_citation_discipline]] — Fletcher & Rossing not PDF-extracted (textbook not on arXiv/PMC); flagged honestly in §5
  • [[feedback_science_is_ssot_not_project]] — Helmholtz/Rayleigh/Fletcher canonical SSoT; closed-form analytical formulae throughout

§8 Fermatas for conductor

  1. Stance candidate: "Kepler equation IS phase modulation at small eccentricity" (Anomaly A1). Mathematical content: ν = M + 2ε·sin(M) + O(ε²) ⟺ FM(M, β=ε) at lowest order. Unifies Class K (pin-slot) with FM synthesis substrate (Chowning 1973). Author as new identity-discipline stance? Or absorb into [[user_stance_kepler_shape_universal]] as explicit substrate instance? User direction required — substantive new identity claim per [[user_stance_identity_not_implementation_discipline]] umbrella.

  2. Spike #41 framing for real-audio cascade-β: real public-domain piano / drum recordings might show dual-rate decay missing from idealised single-exp models. Numerical-resource follow-up.

  3. K-shape in DEEP MUSIC theory (vs acoustic-physics layer): music theory has its own periodic structures (circle of fifths = Class I cyclic group; tonic-dominant = group representations; voice-leading = Class L on note-graph). K/L analysis at music-theory layer rather than acoustic-physics layer would be a complementary spike.

  4. Notebook placement: findings are research-spike-finding-level, not framework-affecting per Spike #38 precedent. The 14-class vocabulary stays at 14; [[user_stance_kepler_shape_universal]] correctly holds. No new stance authored from this spike alone; the Kepler-EOC ≡ FM-small-β identity (Anomaly A1) is the candidate awaiting user direction.

§9 Artifacts

Total: 119 NDJSON records across 7 files; 6 analysis scripts.

End of spike artifact.