MOODIFY
EVIDENCE
Every public claim on this site carries one of four maturity states — Concept, Experimental, Verified, Human-reviewed — plus method, scope, and limitation. Anything here is reproducible from a stable reference where public release permits.
PUBLISH-SAFE ITEMS
52/52 metrics identical across operating systems and Python versions under the frozen scan profile. Method: MFY-WSE-SCAN-PROFILE-001. Limitation: profile-scoped; any profile change requires a new version.
Verified · MFY-WSE-SCAN-PROFILE-001 · 2026-08Deterministic technical ranking within a validated, versioned scope (MFY-ALGORITHMIC-REVIEW-001). Out-of-scope or unresolved perceptual cases escalate to human review or close as inconclusive. Limitation: technical ranking is not artistic quality or copyright conclusion.
Human-reviewed · scope boundary · 2026-08Ten injected failures across upload, queue, and verification paths: eight passed with fail-closed evidence; two findings recorded with owners. Limitation: injected scenarios are not an exhaustive failure catalogue.
Verified · 2026-08Ten complete tracks processed with zero out-of-memory and zero deferred jobs on a 2C/2G node; queue survives restarts. Limitation: single-instance deployment; memory policy is under review.
Verified · 2026-08Sixteen operators across time, spectral, phase, cepstral, modulation, and topological representations accepted as experimental research. Limitation: research-grade thresholds are not production calibrations before September.
Experimental · 2026-08Not published: internal file paths, private audio, prompts, unreviewed judgment logs, or experimental scores presented as rankings.