EVERYTHING RINGS is a local-first acoustic instrument for estimating the audible resonances of struck physical objects, reconstructing those resonances, and turning them into playable instruments.
5
stars
155
commits
TypeScript
primary language
Aug 25, 2026
updated
A local-first acoustic instrument that turns one physical strike into an interpretable resonance model, a playable instrument, and a shareable acoustic artifact.
Open the live instrument → · Product principles · Science & DSP · Full architecture
A glass, bowl, bell, mug, plate, tool, or any other safe object has an audible ring shaped by the transient you record. EVERYTHING RINGS captures that strike in the browser, isolates the ringdown, estimates the resonant structure supported by the recording, reconstructs it, and turns the result into something you can see, hear, play, save, compare, and share.
The first-use loop is intentionally simple:
Underneath that interaction is a much stricter system: deterministic DSP, native-rate acquisition, explicit quality gates, content-addressed provenance, frozen empirical contracts, synthetic qualification, research-only identity/material models, and an emerging Resonance Atlas architecture.
Privacy by design: raw microphone PCM stays local under the current product contract. Local history stores fingerprints, and Acoustic Capsules carry bounded fingerprint data in the URL fragment—not recordings.
| Surface | What it gives you |
|---|---|
| Acoustic DNA | A deterministic visual identity for one measured fingerprint |
| Resonance Microscope | Per-mode frequencies, decay, confidence, and isolated audition |
| Ringdown Lens | A time-domain view of fitted modal envelopes |
| CAPTURE ↔ MODEL | Local comparison between the isolated ringdown and modal reconstruction |
| Playable instrument | Chromatic modal synthesis that preserves the measured frequency ratios and decay constants |
| Saved captures | Fingerprint-only local history that can be reopened without microphone access |
| Resonance Diff | Descriptive comparison of two capture observations without an identity verdict |
| Acoustic Capsules | Fingerprint-only interactive share links: hear it, play it, then try your own |
| Acoustic Story | A self-contained fingerprint-derived sound-and-motion artifact |
flowchart LR
A[Physical strike] --> B[Local microphone capture]
B --> C[Quality gate + ringdown isolation]
C --> D[er-dsp-2 fingerprint]
D --> E[Reveal]
D --> F[Modal reconstruction]
D --> G[Playable instrument]
D --> H[Local history + Capsules]
D --> I[Immutable measurement]
I --> J[Versioned derivations]
J --> K[Repeated-observation acoustic object model]
K --> L[Sonic Twin research]
K --> M[Spatial modal field]
I --> N[Resonance Atlas contracts]
L --> N
M --> N
The architectural rule is simple and important:
A capture is one observation. A fingerprint is one analysis of that observation. Neither is automatically the physical object.
That distinction is what allows the repository to grow from a playful instrument into serious measurement, retrieval, and corpus infrastructure without turning a convenient hash into an unsupported scientific claim.
The evidence-eligible baseline is explicit and inspectable. The canonical er-dsp-2 path estimates peaks, tracks persistent modes, fits decay, computes confidence, and emits a versioned fingerprint. Research estimators live on separate non-evidence paths until a future algorithm earns promotion.
Microphone PCM is not required for sharing, saved-capture playback, Resonance Diff, Twin demos, or Atlas contracts. A future network service must justify any remotely stored data instead of making upload the default architecture.
The validation layer content-addresses immutable measurements independently from renderers, instruments, similarity systems, visualizations, embeddings, and Atlas artifacts. Multi-source derivation records and source links let downstream research evolve without rewriting measurement history.
Capabilities move through an explicit ladder:
prototype → software-qualified → synthetic-qualified → physical-research → held-out-validated → product-eligible → released
A green workflow cannot promote a synthetic result into physical evidence. The claim registry records the strongest proposition currently earned for each capability.
| Capability | Current maturity | Scope |
|---|---|---|
| Consumer STRIKE → REVEAL → HEAR → PLAY | software-qualified | Browser and synthetic qualification |
Resonance estimation (er-dsp-2) | software-qualified | Estimated audible resonances supported by a recorded transient |
| Modal reconstruction + playable instrument | software-qualified | Deterministic fingerprint-derived audio |
| Sonic Twin retrieval / verification | synthetic-qualified | Named digital populations only |
| Spatial modal-field prediction | software-qualified | Predicted fingerprints; never measurement evidence |
| Material inference | prototype / research | No calibrated physical material claim |
| Measurement + derivation provenance | software-qualified | Content-addressed V1 measurements and V2 multi-source derivations |
| Atlas V1/V2 integrity contracts | software-qualified | Records, snapshots, public observations, specimen records, membership assertions |
| Atlas client + local file-backed HTTP API | software-contract qualified | Local/durable software path; not a public production network |
| Empirical Gate A2 / B / C | not yet run | Frozen physical/perceptual sequence remains separate |
The authoritative wording and exclusions live in docs/CLAIM_REGISTRY.md.
The post-freeze software line already contains foundations for the larger research program:
AcousticObjectModelV1 estimates with frequency/decay/amplitude dispersion and observation support;match / nonmatch / abstain;The interactive research surfaces are deliberately labeled synthetic:
?twin=1 — Sonic Twin software lab;?atlas=1 — local Resonance Atlas software lab;?lab=1 — validation/review lab.They are research and qualification surfaces, not product claims about physical identity.
EVERYTHING-RINGS/
├── apps/
│ ├── web/ # consumer instrument + empirical/research surfaces
│ └── atlas-api/ # local/durable Atlas HTTP service contracts
├── packages/
│ ├── acquisition/ # microphone acquisition + capture quality
│ ├── dsp/ # canonical DSP + research estimators
│ ├── fingerprint/ # recurrence, object models, Twin, calibration, material/spatial research
│ ├── synth/ # modal reconstruction
│ ├── instrument/ # playable modal instrument semantics
│ ├── visual/ # fingerprint-derived visual artifacts
│ ├── validation/ # evidence, gates, provenance, Atlas, assurance
│ ├── atlas-client/ # typed Atlas client
│ ├── er-cli/ # independent verification CLI
│ └── fixtures/ # deterministic test fixtures
├── docs/ # product, science, evidence, claims, release and takeover specs
└── scripts/ # browser, lifecycle and end-to-end qualification
Requires Node.js 22 and the repository-declared pnpm version.
corepack enable
pnpm install --frozen-lockfile
pnpm typecheck
pnpm test
pnpm build
pnpm --filter @everything-rings/web dev
The root web app is the consumer experience. For local research surfaces, append ?lab=1, ?twin=1, or ?atlas=1 to the local URL.
pnpm --filter @everything-rings/er-cli test
pnpm --filter @everything-rings/er-cli exec node ./bin/er.mjs hash path/to/file.json
pnpm --filter @everything-rings/er-cli exec node ./bin/er.mjs verify evidence path/to/evidence.json
pnpm --filter @everything-rings/er-cli exec node ./bin/er.mjs verify measurement path/to/measurement.json
pnpm --filter @everything-rings/er-cli exec node ./bin/er.mjs verify derivation path/to/derivation.json
The repository intentionally separates two lines of work:
main is the frozen empirical v8 authority for the current physical validation sequence. Its exact revision is intentionally preserved.post-freeze-development carries product, research, provenance, Atlas, tooling, and future-system work.Software qualification covers deterministic execution, integrity, synthetic acoustic recovery, browser flows, and named digital-population behavior. It does not by itself establish microphone-transducer validity, human perceptual fidelity, physical-object identity, or material identity.
The empirical sequence remains separately preregistered and version-bound. See Gate A, Gate B, Gate C, validation evidence, and the release checklist.
The current evidence-eligible interpretation is:
Estimated audible resonances supported by each recorded transient.
The project does not currently promote a fingerprint into complete structural eigenmodes, a signature into canonical physical-object identity, a research material label into physical material identification, a reconstruction into calibrated loudness, or a digital-twin result into physical validation.
Research code may explore those directions only behind explicit version, population, provenance, uncertainty, and maturity boundaries.
| Start here | Purpose |
|---|---|
docs/PRODUCT.md | Product principles and interaction doctrine |
docs/DSP.md | Canonical signal model and DSP contract |
docs/FULL_VISION_SOFTWARE.md | Long-term software architecture |
docs/ENGINEERING_TAKEOVER_AND_FULL_VISION_SPEC.md | Full engineering/research execution specification |
docs/CLAIM_REGISTRY.md | Capability maturity and exact claim boundaries |
docs/DIGITAL_TWIN_QUALIFICATION.md | Synthetic qualification semantics |
docs/VALIDATION.md | Validation architecture |
docs/RELEASE_CHECKLIST.md | Software release and assurance checklist |
CONTRIBUTING.md | Contribution and change-class rules |
SECURITY.md | Security reporting and sensitive-artifact boundaries |
Contributions are welcome when they preserve the separation between consumer behavior, research diagnostics, and evidence-eligible empirical contracts. Read CONTRIBUTING.md before changing acquisition, DSP, evidence schemas, gate semantics, provenance, or public claim copy.
Software in this repository is released under the MIT License. Dataset, model, and future physical-corpus artifacts may carry separate terms when explicitly stated.
The world is full of things that ring.
This is an instrument for discovering them.
155 commits
Hacker News (1)
TypeScript
90.7%
JavaScript
6.1%
CSS
2.9%
EVERYTHING RINGS is a local-first acoustic instrument for estimating the audible resonances of struck physical objects, reconstructing those resonances, and turning them into playable instruments.
5
stars
155
commits
TypeScript
primary language
Aug 25, 2026
updated
A local-first acoustic instrument that turns one physical strike into an interpretable resonance model, a playable instrument, and a shareable acoustic artifact.
Open the live instrument → · Product principles · Science & DSP · Full architecture
A glass, bowl, bell, mug, plate, tool, or any other safe object has an audible ring shaped by the transient you record. EVERYTHING RINGS captures that strike in the browser, isolates the ringdown, estimates the resonant structure supported by the recording, reconstructs it, and turns the result into something you can see, hear, play, save, compare, and share.
The first-use loop is intentionally simple:
Underneath that interaction is a much stricter system: deterministic DSP, native-rate acquisition, explicit quality gates, content-addressed provenance, frozen empirical contracts, synthetic qualification, research-only identity/material models, and an emerging Resonance Atlas architecture.
Privacy by design: raw microphone PCM stays local under the current product contract. Local history stores fingerprints, and Acoustic Capsules carry bounded fingerprint data in the URL fragment—not recordings.
| Surface | What it gives you |
|---|---|
| Acoustic DNA | A deterministic visual identity for one measured fingerprint |
| Resonance Microscope | Per-mode frequencies, decay, confidence, and isolated audition |
| Ringdown Lens | A time-domain view of fitted modal envelopes |
| CAPTURE ↔ MODEL | Local comparison between the isolated ringdown and modal reconstruction |
| Playable instrument | Chromatic modal synthesis that preserves the measured frequency ratios and decay constants |
| Saved captures | Fingerprint-only local history that can be reopened without microphone access |
| Resonance Diff | Descriptive comparison of two capture observations without an identity verdict |
| Acoustic Capsules | Fingerprint-only interactive share links: hear it, play it, then try your own |
| Acoustic Story | A self-contained fingerprint-derived sound-and-motion artifact |
flowchart LR
A[Physical strike] --> B[Local microphone capture]
B --> C[Quality gate + ringdown isolation]
C --> D[er-dsp-2 fingerprint]
D --> E[Reveal]
D --> F[Modal reconstruction]
D --> G[Playable instrument]
D --> H[Local history + Capsules]
D --> I[Immutable measurement]
I --> J[Versioned derivations]
J --> K[Repeated-observation acoustic object model]
K --> L[Sonic Twin research]
K --> M[Spatial modal field]
I --> N[Resonance Atlas contracts]
L --> N
M --> N
The architectural rule is simple and important:
A capture is one observation. A fingerprint is one analysis of that observation. Neither is automatically the physical object.
That distinction is what allows the repository to grow from a playful instrument into serious measurement, retrieval, and corpus infrastructure without turning a convenient hash into an unsupported scientific claim.
The evidence-eligible baseline is explicit and inspectable. The canonical er-dsp-2 path estimates peaks, tracks persistent modes, fits decay, computes confidence, and emits a versioned fingerprint. Research estimators live on separate non-evidence paths until a future algorithm earns promotion.
Microphone PCM is not required for sharing, saved-capture playback, Resonance Diff, Twin demos, or Atlas contracts. A future network service must justify any remotely stored data instead of making upload the default architecture.
The validation layer content-addresses immutable measurements independently from renderers, instruments, similarity systems, visualizations, embeddings, and Atlas artifacts. Multi-source derivation records and source links let downstream research evolve without rewriting measurement history.
Capabilities move through an explicit ladder:
prototype → software-qualified → synthetic-qualified → physical-research → held-out-validated → product-eligible → released
A green workflow cannot promote a synthetic result into physical evidence. The claim registry records the strongest proposition currently earned for each capability.
| Capability | Current maturity | Scope |
|---|---|---|
| Consumer STRIKE → REVEAL → HEAR → PLAY | software-qualified | Browser and synthetic qualification |
Resonance estimation (er-dsp-2) | software-qualified | Estimated audible resonances supported by a recorded transient |
| Modal reconstruction + playable instrument | software-qualified | Deterministic fingerprint-derived audio |
| Sonic Twin retrieval / verification | synthetic-qualified | Named digital populations only |
| Spatial modal-field prediction | software-qualified | Predicted fingerprints; never measurement evidence |
| Material inference | prototype / research | No calibrated physical material claim |
| Measurement + derivation provenance | software-qualified | Content-addressed V1 measurements and V2 multi-source derivations |
| Atlas V1/V2 integrity contracts | software-qualified | Records, snapshots, public observations, specimen records, membership assertions |
| Atlas client + local file-backed HTTP API | software-contract qualified | Local/durable software path; not a public production network |
| Empirical Gate A2 / B / C | not yet run | Frozen physical/perceptual sequence remains separate |
The authoritative wording and exclusions live in docs/CLAIM_REGISTRY.md.
The post-freeze software line already contains foundations for the larger research program:
AcousticObjectModelV1 estimates with frequency/decay/amplitude dispersion and observation support;match / nonmatch / abstain;The interactive research surfaces are deliberately labeled synthetic:
?twin=1 — Sonic Twin software lab;?atlas=1 — local Resonance Atlas software lab;?lab=1 — validation/review lab.They are research and qualification surfaces, not product claims about physical identity.
EVERYTHING-RINGS/
├── apps/
│ ├── web/ # consumer instrument + empirical/research surfaces
│ └── atlas-api/ # local/durable Atlas HTTP service contracts
├── packages/
│ ├── acquisition/ # microphone acquisition + capture quality
│ ├── dsp/ # canonical DSP + research estimators
│ ├── fingerprint/ # recurrence, object models, Twin, calibration, material/spatial research
│ ├── synth/ # modal reconstruction
│ ├── instrument/ # playable modal instrument semantics
│ ├── visual/ # fingerprint-derived visual artifacts
│ ├── validation/ # evidence, gates, provenance, Atlas, assurance
│ ├── atlas-client/ # typed Atlas client
│ ├── er-cli/ # independent verification CLI
│ └── fixtures/ # deterministic test fixtures
├── docs/ # product, science, evidence, claims, release and takeover specs
└── scripts/ # browser, lifecycle and end-to-end qualification
Requires Node.js 22 and the repository-declared pnpm version.
corepack enable
pnpm install --frozen-lockfile
pnpm typecheck
pnpm test
pnpm build
pnpm --filter @everything-rings/web dev
The root web app is the consumer experience. For local research surfaces, append ?lab=1, ?twin=1, or ?atlas=1 to the local URL.
pnpm --filter @everything-rings/er-cli test
pnpm --filter @everything-rings/er-cli exec node ./bin/er.mjs hash path/to/file.json
pnpm --filter @everything-rings/er-cli exec node ./bin/er.mjs verify evidence path/to/evidence.json
pnpm --filter @everything-rings/er-cli exec node ./bin/er.mjs verify measurement path/to/measurement.json
pnpm --filter @everything-rings/er-cli exec node ./bin/er.mjs verify derivation path/to/derivation.json
The repository intentionally separates two lines of work:
main is the frozen empirical v8 authority for the current physical validation sequence. Its exact revision is intentionally preserved.post-freeze-development carries product, research, provenance, Atlas, tooling, and future-system work.Software qualification covers deterministic execution, integrity, synthetic acoustic recovery, browser flows, and named digital-population behavior. It does not by itself establish microphone-transducer validity, human perceptual fidelity, physical-object identity, or material identity.
The empirical sequence remains separately preregistered and version-bound. See Gate A, Gate B, Gate C, validation evidence, and the release checklist.
The current evidence-eligible interpretation is:
Estimated audible resonances supported by each recorded transient.
The project does not currently promote a fingerprint into complete structural eigenmodes, a signature into canonical physical-object identity, a research material label into physical material identification, a reconstruction into calibrated loudness, or a digital-twin result into physical validation.
Research code may explore those directions only behind explicit version, population, provenance, uncertainty, and maturity boundaries.
| Start here | Purpose |
|---|---|
docs/PRODUCT.md | Product principles and interaction doctrine |
docs/DSP.md | Canonical signal model and DSP contract |
docs/FULL_VISION_SOFTWARE.md | Long-term software architecture |
docs/ENGINEERING_TAKEOVER_AND_FULL_VISION_SPEC.md | Full engineering/research execution specification |
docs/CLAIM_REGISTRY.md | Capability maturity and exact claim boundaries |
docs/DIGITAL_TWIN_QUALIFICATION.md | Synthetic qualification semantics |
docs/VALIDATION.md | Validation architecture |
docs/RELEASE_CHECKLIST.md | Software release and assurance checklist |
CONTRIBUTING.md | Contribution and change-class rules |
SECURITY.md | Security reporting and sensitive-artifact boundaries |
Contributions are welcome when they preserve the separation between consumer behavior, research diagnostics, and evidence-eligible empirical contracts. Read CONTRIBUTING.md before changing acquisition, DSP, evidence schemas, gate semantics, provenance, or public claim copy.
Software in this repository is released under the MIT License. Dataset, model, and future physical-corpus artifacts may carry separate terms when explicitly stated.
The world is full of things that ring.
This is an instrument for discovering them.
Hacker News (1)
155 commits
TypeScript
90.7%
JavaScript
6.1%
CSS
2.9%