1op Senses · pre-alpha animal perception interface

Every way of seeing the world

1op Senses is an interactive perception layer for animal sensing models. It maps biological sensing modes into EML-inspired model sketches and browser demos across vision, hearing, smell, touch, electroreception, magnetoreception, and infrared. This is pre-alpha product work, not biological proof, hardware validation, medical/veterinary advice, or production autonomy.

Browser demos only
No microphone required
No hardware required
Web Audio oscillator mappings only

Vision

3 species
Live demochain 1

Cat

Vision

Dichromatic — two cones, no red — plus a tapetum lucidum that returns photons to the retina for a second pass.

S(λ) = exp(-((λ - λ_peak)² / (2σ²))) ; tapetum_gain ∈ [1, 130]
~6× human low-light sensitivityOpen demo
Live demochain 1

Mantis shrimp

Vision

16 photoreceptor channels spanning 300–720 nm plus 4 polarization channels. UV-fluorescent patches that humans see as plain coral light up as blue-violet glow in the mantis view.

S_n(λ) = exp(-((λ - λ_n)² / (2σ²)))   for n ∈ {1, …, 16}
16 channels · 3 UV · 4 polarizationOpen demo
Live demochain 1

Bee

Vision

Trichromat with UV — three Gaussian cones at 350, 440, 540 nm replace human RGB. Flowers paint UV-reflective "nectar guides" invisible to us; for the bee, every sunflower has a runway.

S_uv(λ), S_blue(λ), S_green(λ) = exp(-((λ - λ_n)² / (2σ²)))
UV cone at 350 nm · sunflower bullseye visibleOpen demo

Hearing

3 species
Live demochain 0-1

Bat

Hearing

Active sonar. The bat broadcasts an ultrasonic chirp and ranges the world from the round-trip echo time. The world's first FMCW radar — 50 million years old. Five EML files: chirp, echo, Doppler, beam, matched filter.

range(t) = c · Δt / 2 ;   c = 343 m/s
1 mm range · 0.5 m/s velocity recoveredOpen demo
Live demochain 0-1

Dolphin

Hearing

Underwater FMCW sonar — the same matched-filter math the bat uses, but at 1500 m/s in seawater instead of 343 in air. ~4.4× the reach for the same time budget; click rate ramps up as targets close in.

range(t) = c · Δt / 2 ;   c = 1500 m/s
~4.4× bat's reach for the same delayOpen demo
Live demochain 1

Owl

Hearing

Asymmetric ear placement — one higher than the other — turns the inter-aural time delay into a 3D localization vector. Hunts mice in pure darkness, by sound alone.

Δt = d · sin(θ) / c   ⇒   θ = arcsin(c · Δt / d)
1.7° RMS vs human's 42° — pure darkOpen demo

Smell

1 species
Live demochain 0-1

Dog

Smell

300 M olfactory receptors vs the human 6 M, plus a stereo-nostril gradient cue. Trace explosive plume invisible to human noses; legible to dogs across the test domain.

r(c) = c / (c + K_d)   (Michaelis-Menten)
10,000× tighter K_d at trace concentrationsOpen demo

Touch

1 species
Live demochain 0-1

Spider

Touch

The web is the sensor. Standing waves, mechanical bandpass filter, direction-of-arrival from inter-radial delay, prey-vs-wind classifier from frequency content. Passive analog signal processing.

G(f) = 1 / √(1 + ((f² - f₀²) / (f · Δf))²)
PREY +1.00 · WIND -0.18 (web-filtered)Open demo

Electroreception

2 species
Live demochain 0-1

Shark

Electroreceptionno human analog

The ampullae of Lorenzini detect electric fields down to 5 nV/cm. A buried fish is invisible in turbid water but its bioelectric dipole pulls the shark in to within 30 cm.

|E| = p · |cos θ| / (4πε r³)
30 cm detection range · matches Kalmijn 1971Open demo
Live demochain 1

Electric eel

Electroreceptionno human analog

Active electroreception — the eel generates its own dipole field and reads how nearby objects distort it. Conductive prey raise |E| at the skin electrodes; insulating rocks lower it. Inverse problem: object class from the polarity of the perturbation pattern. Impedance imaging.

ΔE / E ≈ Γ · (a / r)³ ,  Γ = (σ_obj − σ_w) / (σ_obj + 2σ_w)
polarity-only classifier · prey vs rock at any rangeOpen demo

Magnetoreception

1 species
Live demochain 1

Pigeon

Magnetoreceptionno human analog

Cryptochrome proteins in the retina form radical pairs whose recombination rate depends on Earth's magnetic field direction. Quantum biology you can fly with.

Φ(B,θ) = ⅓ + (2/15)κ(3·cos²(θ) - 1)
|B| 30–60 µT · inclination 0–90° mapOpen demo

Infrared

1 species
Live demochain 0-1

Pit viper

Infraredno human analog

The loreal pit organ is a biological bolometer — a thin membrane suspended in air, heated by IR radiation, sensed by mechanoreceptors. 0.003 °C resolution.

B(λ, T) = (2hc² / λ⁵) · 1 / (exp(hc / λkT) - 1)   (Planck)
37 °C target = 5,667× thresholdOpen demo

Local sensory instrument prototype

Animal senses explorer

A browser-only instrument for mapping biological sensing models into visual and audio placeholders. It uses local data, Web Audio oscillator tones, no microphone, no hardware, and no external network calls.

Boundary strip: not biological proof, not a certified sensor model, not hardware validation, not production autonomy, not medical/veterinary advice, and not a public theorem/proof/open-problem claim.
Visual mappingColor, brightness, UV, and channel-density preview.blue-green biased image with boosted low-light contrast

Active species / modality

Cat vision

Dichromatic vision with low-light gain from the tapetum lucidum.

Modalityvision
Waveformsine
Base Hz740
Stateidle
Audio mappingSoft high shimmer tone for brightness and channel density.Rhythm: soft high shimmerBrowser Web Audio only; no microphone, audio file, hardware, or network call.

Perception legend

Input signalambient wavelength distribution and low-light photon count
Output perceptblue-green biased image with boosted low-light contrast
Model sketchTwo broad cone-response curves plus bounded luminance gain.
EML placeholdercat_vision(signal) -> dichromat_channels + tapetum_gain

Public candidate · local browser prototype

Operator Senses

Explore simple reciprocal operator kernels as visual and browser-audio mappings. Operator Senses is a pre-alpha sensory-math playground for simple reciprocal operator kernels. It is not a theorem prover, not a physics claim, and not a safety system.

Boundary strip: pre-alpha sensory math toy kernels. Not a theorem/proof/open-problem claim. Not physics or holography proof. Not certified safety. Not production controller evidence. Not a Mathlib replacement. No microphone. No hardware action.
Gallery kernels10
Factory kernels25
Audio runtimebrowser Web Audio
Visual mappingalpha and beta show reciprocal contraction/expansion. The event curve uses n = x(a*x+b). Singularity guards mark x = 0 and x = -b/a.

y=2*x+3

alpha = x/(2*x+3); beta = (2*x+3)/x; identity sample alpha*beta = 1

a2
b3
alpha limit0.5
beta limit2
active x-5
tone stateidle
Singularity guards0, -1.5Audio maps alpha as a lower contraction tone, beta as expansion modulation, and the identity relation as a stable short drone.
alpha / beta / identity samples
x = -5alpha: 0.714
beta: 1.4
identity: 1
event n: 35
x = -2alpha: 2
beta: 0.5
identity: 1
event n: 2
x = -1alpha: -1
beta: -1
identity: 1
event n: -1
x = -0.5alpha: -0.25
beta: -4
identity: 1
event n: -1
x = 0.5alpha: 0.125
beta: 8
identity: 1
event n: 2
x = 1alpha: 0.2
beta: 5
identity: 1
event n: 5
x = 2alpha: 0.286
beta: 3.5
identity: 1
event n: 14
x = 5alpha: 0.385
beta: 2.6
identity: 1
event n: 65

Local public candidate · existing EML

EML Playground

Choose a small existing EML expression, inspect a toy tree, map it into visuals and browser Web Audio, and preview a local private evidence JSON record. This does not create a new EML language.

Boundary strip: pre-alpha EML sensory playground. Uses existing EML concepts. Not a theorem/proof/open-problem claim. Not a Mathlib replacement. Not certified safety. Not production controller evidence. No microphone. No hardware action.
Examples5
Uses existing EMLtrue
Audio runtimebrowser Web Audio
Visual mappingTree depth controls band height. Chain estimate controls pulse density. Related operator mappings are labeled as related toy mappings.

eml(1, 1)

balanced two-input node with a compact center pulse. short stable tone with a light rise and fall.

Depth1
Chain1
Active nodeeml
Tone stateidle

Copy evidence state: idle

Toy expression tree
eml
literal: 1
literal: 1
Local/private evidence preview
{
  "capcard_public_marketplace": false,
  "certified_safety_claim": false,
  "depth_estimate": 1,
  "example_id": "eml_constant_balance",
  "expression": "eml(1, 1)",
  "huggingface_upload_performed": false,
  "not_claimed": [
    "not a theorem proof",
    "not a Mathlib replacement",
    "not certified safety",
    "not production controller evidence"
  ],
  "petal_api_upload_performed": false,
  "production_controller_claim": false,
  "public_claim": false,
  "senses_mapping": {
    "visual": "centered pulse",
    "audio": "stable oscillator tone",
    "interaction": "single-step constant example"
  },
  "theorem_proof_claim": false
}

Same language for the chip and the cat.

The direction is one model vocabulary that can describe biological sensing, simulation, visualization, and future hardware targets. Today this is local product groundwork: interactive sensory interfaces, EML-shaped model sketches, and carefully scoped artifacts rather than a claim of certified autonomy or hardware-validated deployment.