Master Evidence Table
Five distinct acoustic signatures have been identified across eleven independent recordings using four complementary methods: TDOA multilateration, 3D delay-and-sum beamforming, inter-camera phase coherence analysis, and continuous wavelet decomposition. Each signature has been independently cross-validated against the Ken Linke 21-recorder crack-boom dataset, Canon XA55 professional PCM audio, and two independent AudioFreq reports.
| Signature | Timing (re: rifle fire) | Distance | Key Evidence | Cross-Validated |
|---|---|---|---|---|
| Mach Cone | t = 0 (reference) | 127m rooftop | N-wave at 11 cameras · 733 sub-200µs zero-crossings (Canon CH1) · 20.8µs fastest · GN formula Vb median 2,241 fps (21 recorders) | ✓ All Sources |
| Stage Detonation | +191ms at tent +321ms at Canon |
Tent / collar | 104µs rise time · 4.9× faster than PETN · 656Hz spectral centroid vs 2,429Hz muzzle blast · FR05m: no local phase collapse | ✓ All Sources |
| Chest Resonance | +136ms peak (IMG_6368) | Victim thorax | 53Hz, 80Hz, 133Hz modes · +14.1dB lung Helmholtz over chest wall · Wavelet +17dB at 52Hz · 47–58Hz bounded prediction at FRC-scale | ✓ All Sources |
| Muzzle Blast | +231ms mean (10 cams) +202ms Canon · +203ms FR05m |
127m rooftop | 21 recorders 109–195ms crack-boom · Beamforming hotspot ≠ tent · FR05m boom at +203ms · Canon off-axis 25ms confirms Mach cone geometry | ✓ All Sources |
| 4940Hz Strouhal | +0.5ms (Mach cone onset) | 3.3m from van | Strouhal: f = 0.2×188÷0.00762 = 4,934Hz · 92,726σ amplitude · TDOA residual 0.000024 · L/R coherence 0.199→1.000 near→far-field transition | ✓ All Sources |
Muzzle Blast Cross-Validation
Four independent datasets all converge on a 127m shooter. The opposition report (AudioFreq)
claimed a 130m shooter would produce a 214–235ms crack-boom gap, rendering Losee incompatible
with the data. That claim contains a geometric error: the formula
Δt = d(1/c − 1/v) is derived for an observer co-located with the target.
The FR05m microphone is 5 meters from Charlie Kirk. Correcting for this offset
brings the predicted interval to ~219ms — with the measured 203ms falling
within measurement uncertainty of a 127m shooter.
Ken Linke — 21 Recorders
Nearest recorders (~5m from CK): 188–194ms
Farthest recorders (~128m from muzzle): 109–131ms
GN formula Vb: median 2,241 fps (supersonic recorders)
All consistent with 127m shooter at every recorder position.
Canon XA55 — +202ms
Canon is off-axis from the shooter-to-CK line by its GPS position. Collinear recorders at equivalent b_3d distance (357ft) measure 174–177ms. Canon measures 202ms — 25ms longer due to Mach cone retarded-time geometry for off-axis receivers. This discrepancy is itself confirmation of moving-source Mach cone physics.
FR05m Mic — +203ms (AudioFreq Data)
AudioFreq identifies boom at 8.487s, crack at 8.284s
→ interval 203ms. Our nearest recorders at 5m from CK: 188–194ms.
FR05m is also ~5m from CK → 203ms is within 10ms. Consistent.
AudioFreq's own data validates the 127m hypothesis.
AudioFreq Prediction — Formula Error
Claims 130m → 214–235ms. Uses Δt = d(1/c − 1/v) with observer AT Charlie.
Corrected formula (5m mic offset): ~219ms predicted.
FR05m measured: 203ms. Difference = 16ms within measurement noise.
130m is consistent with all measured data.
All four datasets — Ken Linke 21 recorders, Canon XA55, FR05m microphone, and AudioFreq's own report — are consistent with a 127m shooter. AudioFreq's claim that 130m is "incompatible" with the data is based on a formula that assumes the observer is standing at Charlie Kirk's position. When corrected for the actual microphone position, the 130m hypothesis is confirmed, not excluded.
Stage Detonation — Multi-Source Confirmation
The stage detonation is confirmed by four independent lines of evidence: the Canon XA55 PCM recording resolving it as the strongest low-frequency event at +321ms, the 10-camera 104µs rise time comparison against confirmed PETN detonations, spectral separation from the muzzle blast, and the AudioFreq FR05m phase continuity finding.
Mach cone immediately begins propagating. Bullet at 880 m/s.
Mach cone sweeps stage. 4940Hz Strouhal tone initiates at van (+0.5ms offset).
43ms after bullet arrival. Electronics chain <10ms; remainder consistent with LiPo thermal cascade (150–500ms published) or pyrotechnic delay.
35ms AFTER detonation. The detonation cannot be the muzzle blast.
Tent-to-Canon travel = 134ms. Back-calculated: detonation at tent = +187ms. Exact match.
At 13.mp4 (4m from tent). 4.9× faster than confirmed PETN (505µs avg Lebanon pager attacks). 48× faster than LiPo thermal runaway (1–10ms).
AudioFreq's independent FR05m analysis: no clipping, no phase collapse at the 5m microphone. Detonation was NOT at FR05m → must be spatially separated → consistent with tent/collar origin.
Detonation centroid: 656Hz. Muzzle blast centroid: 2,429Hz. 3.7× spectral separation confirms two distinct acoustic sources, not one.
AudioFreq's FR05m report concludes "nothing exploded near the microphone." This is not contradictory — it is corroborating. A directional shaped charge deposits energy inward (jet) and along the collar axis. The outward radial component reaching the 5m FR05m mic would arrive as a normally propagating pressure wave, indistinguishable from any other acoustic source at that distance. The absence of local overload at FR05m confirms the detonation was directional and spatially separated from the mic — exactly consistent with the tent/collar origin, not a contradiction of it.
4940Hz Strouhal Tone — Van Localization Confirmed
The 4940Hz signature is confirmed by four independent methods and is entirely unaddressed by the AudioFreq reports. The near-to-far-field phase coherence transition across camera distances is the definitive proof that this is a real acoustic point source, not an electronic artifact, resonator, or codec artifact.
Electronic artifacts, PA feedback, EMP, codec compression, and resonator artifacts produce identical signals on both channels regardless of distance — they cannot produce a distance-dependent coherence transition. Only a real acoustic signal propagating through air from a spatially fixed point source produces this pattern: low coherence (extreme wavefront curvature) near the source, high coherence (plane wave) far from the source. The transition from 0.199 at 9.2m to 1.000 at 30.4m is the acoustic fingerprint of a real, localized sound source — at the van position.
Strouhal Equation
f = St × v / D = 0.2 × 188 m/s ÷ 0.00762m = 4,934 Hz
Measured: 4,940 Hz. Gas venting through a .30-caliber bullet hole in 20%
ballistic gelatin at Mach 0.55. Confirmed by xAI/Grok aeroacoustic analysis.
92,726σ Amplitude
Standard deviations above baseline. Statistically eliminates the entire class of jet-edge resonators, cylindrical resonators, hysteresis systems, and stochastic acoustic processes — all of which produce peaks with gradual onset and amplitude variance. A 92,726σ impulse requires an instantaneous, high-energy source. There is no laboratory resonator that produces this.
TDOA Localization
4-camera clean set. Converges on (−2.04, −1.40)m from tent. Distance from van: 3.29m. Multilateration residual: 0.000024 — essentially a perfect fit. The AudioFreq reports do not address this source. A van at 3.3m from the tent would produce a ~19ms reflection in AudioFreq's data that he does not identify — a significant gap in his analysis.
Cavitation Collapse Pulses
Periodic 300–3kHz broadband impulses at 5.4ms mean interval detected at 2.MOV (closest camera to van). Pattern matches ballistic gel temporary cavity oscillation — the gel "breathing" as it collapses and re-expands after bullet penetration. A battery thermal runaway does not produce periodic cavitation signatures.
Opposition Reports — Errors and Useful Data
Two independent reports were produced by AudioFreq: a TDOA analysis (September 2025) concluding the shooter was in the "south courtyard at ~50–95m," and a FR05m microphone analysis (January 2026) concluding no explosive event occurred near the 5m microphone. Both contain valuable usable data; both also contain critical methodological errors.
Critical Errors
The formula Δt = d(1/c − 1/v) is valid only when the observer
is at the target. The FR05m mic is 5m from Charlie. Corrected prediction for
130m shooter: ~219ms. Measured: 203ms.
130m is consistent.
Claims 130m → 214–235ms. Ken Linke 21-recorder dataset shows equivalent distances produce 109–131ms. AudioFreq's formula overshoots real measurements by ~85ms — the geometric error accounts for the entire discrepancy.
All TDoA reflection analysis applies a stationary point source model to the ballistic crack. The crack is a moving supersonic source. Stationary model applied to crack arrivals across 11 GPS cameras: converges to 192 meters underground. The "south courtyard" localization is a spurious artifact of wrong physics.
Uses ±80ms human reaction window on an 80–160ms signal. This introduces ±50% uncertainty into the distance estimate. The resulting "50–95m" range already contains 130m within one standard deviation.
Independently Useful Data
All reflection distance calculations (38m, 76m, 87m, 103m wall offsets) independently verified correct. These represent real reflective surfaces in the UCCU Center courtyard. The 103m return from Losee is genuine.
AudioFreq himself concludes Losee appears only as a late higher-order echo across all five microphones — not as the direct source. This is fully consistent with our 127m n_wave KML position for the shooter.
No clipping, no phase collapse, no local overload at the 5m microphone. Confirms the stage detonation was NOT omnidirectional. Consistent with a directional shaped charge at the tent/collar origin.
8.284s crack · 8.487s boom (+203ms) provides a verified sync anchor cross-validating Ken Linke nearest recorders (188–194ms) within 10ms. The third event at +554ms remains unexplained and requires sync investigation.
The Definitive Test — 192 Meters Underground
The most powerful single result of this integrated analysis is the stationary source model applied to the crack arrival times across eleven GPS-positioned cameras. If the crack were a stationary point source — as AudioFreq's reflection model assumes — the least-squares optimizer should converge on a real, physical location. It does not.
RMS residual. Fitted source location: (−227m, 0, −192m). 192 meters underground. Physically impossible. No rock strata fired a rifle.
RMS residual. Fitted source location: exact n_wave GPS coordinates from KML. Methodology validated. The boom IS a stationary source. Apply same model to crack: it fails.
His reflection-based TDoA model treats the crack as a fixed point source. The optimizer finds the least-bad stationary fit to a moving source. "South courtyard" is a mathematical artifact, not physical evidence.
The boom (muzzle blast) is a known stationary source. Applying the stationary model to boom arrivals: 0.00ms RMS, fits exact GPS coordinates. This confirms the methodology works when applied to a stationary source. Applying the identical methodology to crack arrivals: 39ms RMS, source 192 meters underground. This confirms the crack is NOT a stationary source. The methodology validates itself — and falsifies AudioFreq's fundamental assumption in a single experiment.
Download and Verify
Eleven independent recordings. Five distinct acoustic signatures from three spatially
separated locations. A detonation at the tent spectrally distinct from the rifle.
Audio-visual delay of 0–2 frames, not 11. Radial motion from the transmitter.
A necklace launched at 47 mph. The opposition's own reports confirm the muzzle blast
timing is consistent with 127m, confirm no local explosive at the 5m mic,
and confirm Losee is a distant reflector only.
Reasonable doubt does not require proving what happened.
It requires showing that a single rifle shot from 127 meters cannot account
for the physical evidence. It cannot. Not even close.