SPECTRALAUDIT REPORT

Version: 0.3.0

INTERPRETATION
Origin inconclusive: spectral characteristics do not establish provenance.

METADATA
path: examples/assets/synthetic_spectral_comb.wav
sha256: add58b1f1516f4f2993fd5f5c3a81a1a434159370cca9a78b78a5385220e1178
samplerate_hz: 44100
duration_s: 6.0
frames: 264600
channels: 1
subtype: PCM_16
peak_amplitude: 0.32391357421875
rms_amplitude: 0.05171883853742848
clipped_sample_fraction: 0.0
sample_peak_only: true
analysis_signal: mono

FINDINGS
  SYNTHETIC_PATTERNS
    method: Welch log-PSD; Gaussian envelope removal; temporal median; residual peak spacing and autocorrelation
    band_hz: [5000.0, 16000.0]
    sample_rate_hz: 44100
    not_generator_specific: true
    origin_attribution: not_supported
    research_context: Architecture-linked spectral artifacts studied in Suno/Udio; not a Suno detector
    available: true
    analyzed_window_count: 2
    total_full_windows: 2
    peak_frequencies_hz: [6099.28, 6540.71, 6982.14, 7418.19, 7859.62, 8301.05, 8742.48, 9178.53, 9619.96, 10061.39, 10497.44, 10938.87, 11380.3, 11821.73, 12257.78, 12699.21, 13140.64, 13582.07, 14018.12, 14459.55, 14900.98]
    peak_count: 21
    estimated_spacing_hz: 441.4306640625
    relative_spacing_mad: 0.0
    regularity: 1.0
    mean_peak_residual_db: 22.887399091806255
    mean_peak_prominence_db: 25.027407687920448
    persistence_fraction: 1.0
    autocorrelation_spacing_hz: 441.4306640625
    autocorrelation_strength: 0.906786830345
    candidate_pattern: true
    interpretation: Persistent periodic high-frequency structure found; may be compatible with architecture-linked artifacts but is not unique to AI music or Suno.

LIMITATIONS
- Frequency-domain heuristics are not verified AI provenance indicators.
- Lossy encoding, mastering, harmonic instruments, and source bandwidth can confound measurements.
- Whole-file read may require substantial RAM; no streaming interface.
- Cutoff and periodicity thresholds are experimental and not calibrated on real-world corpora.
- Suno and other generators cannot be identified reliably from a single spectral test.

SCIENTIFIC NOTE
SpectralAudit measures spectral patterns, not AI authorship. Its findings do not constitute proof that a recording was generated by artificial intelligence.
