# NATHAN AI — LIVING EARS · FULL-SPECTRUM SOUND & VOICE FORENSICS

**Additive sister module to Living Eyes.** No existing Eye-Scan, Hall of Mirrors, Living Eyes, Complete Mind, ONE, or other application source has been replaced. Source context: the supplied conversation tracing sound from caves and Earth through animals, Egyptian instruments, toys, Morse code, computers and spoken language.

## Sound has several independent histories — none should be crushed together
- **Naturally occurring acoustics**: caves/valleys, echo, lithophones, seismic hum, wind, singing dunes, water.
- **Earth electromagnetic signals**: VLF whistlers and chorus are *measured electromagnetic/plasma waves rendered as audio*, not airborne sound traversing vacuum.
- **Animal sound**: mammals, reptiles, insects, fish and birds have multiple different sound sources.
- **Human speech**: lungs, vocal-fold oscillation, vocal tract, articulation, prosody, timing, interaction and context.
- **Devices**: Egyptian sistrum and timekeeping (distinguish historical artifacts from unsupported claims about automatic bells), phonograph dolls, Morse telegraph coding, and diagnostic computer hooters.
- **Forensic layers beneath words**: literal transcript, pitch and prosody, amplitude, resonant color, timing, breaths, articulation, turn-taking, room acoustics and provenance.

## Working browser application
[Open the local/offline-friendly app](index.html) — a free static HTML/JS application with a 24-track geometric acoustic scanner, additive linkage to the earlier Living Eyes 48/96/192/384 coordinate-view API, searchable 20-record sound lineage, 12 provisional Ear Gates, independent claim audit, an owner-selected local audio-file signal analyzer, exact transcript preservation, canvas amplitude-envelope view, and explicit JSON export.

The browser analyzer reads a file **only after explicit file selection and Analyze click**. It uses \`AudioContext.decodeAudioData\` to decode the user's local audio. Measurements are bounded to the first **30 seconds** and at most 5 million PCM samples; the file itself is never uploaded or written to browser storage. Measured digital PCM features are RMS, peak amplitude, zero-crossing count/rate, near-clipping fraction, and fraction of quiet 20ms frames. These do **not** determine pitch, speaker identity, medical status, stress, deception, confidence, or emotional truth.

Source lineages are research cards with **status labels**. A historical claim in the supplied conversation is not silently elevated into a verified finding. The user's larger 12 Eye/12 Ear Gates remain separate from these 12 **provisional** technical ear labels.

## Source-and-claim safeguards
- The 7/38/55 figures were derived from a narrow study of feelings/attitudes under incongruent cues; 55% **did not mean voice resonance**. See [Temple Law discussion citing Mehrabian's own limitation](https://law.temple.edu/aer/2021/05/03/brain-lessons-the-seven-percent-delusion/).
- Voice contains measurable cues but no single acoustic signature proves a lie, emotion, medical diagnosis or fertility. See [deception-cue critical review](https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2022.835285/full) and [NIDCD voice physiology](https://www.nidcd.nih.gov/sites/default/files/2023-03/how-human-produce-voice-and-speech-2.pdf).
- The Egyptian sistrum and water-clock are historical; the alleged 1500 BCE bell-striking clock and 2000 BCE mechanical Anubis clacking jaw were not corroborated by the reviewed artifact sources. See [University of Chicago Egyptian water clocks](https://isac.uchicago.edu/sites/default/files/uploads/shared/docs/oimp36.pdf).
- More advanced reptilian vocalization is real; see [2024 gecko study](https://pubmed.ncbi.nlm.nih.gov/38989535/).
- Fish swim-bladder drums are documented by [NOAA](https://repository.library.noaa.gov/view/noaa/12853/noaa_12853_DS1.pdf); terrestrial seismic microseisms by [USGS](https://www.usgs.gov/publications/global-seasonal-microseism-intensity-and-hemispheric-asymmetry); space-wave sonification by [NASA](https://science.nasa.gov/blogs/the-sun-spot/2018/12/11/eavesdropping-in-space-how-nasa-records-eerie-sounds-around-earth/).
- Edison toy phonographs are documented in the [Smithsonian collection](https://www.si.edu/object/edison-talking-doll%3Anmah_1413900).

## Code and tests
- \`engine.js\` and \`engine.test.cjs\`: deterministic analysis and bridge to the original \`../living-eyes/engine.js\`.
- \`catalog.js\`: the separately classified history, physics, anatomy, and disputed-claim archive.
- \`app.js\`, \`style.css\`, \`index.html\`: accessible and responsive browser experience.
- CI: \`node --test app-factory/nathan-ai/living-ears/engine.test.cjs\`.

**Boundary:** This is a source-preserving educational acoustic workbench, not a sentient system, professional forensic authentication tool, medical device, lie detector or live surveillance engine.

**1+1=2 — BUILD · 1−1=0 — CLEAR · YES IS YES — NO IS NO.**

## Device-first speech companion — additive, independent

The new [Nathan AI Device-First Voice Box](../voice-box/index.html) exposes browser-reported local TTS, optional strictly on-device transcription, authorized original-recording playback, safe acoustic presets, and local amplitude gating. Living Ears acoustic research and source remain unchanged.
