Ondova Talk to us
Intelligent sensing for ear-worn devices

Make every device worn on the ear understand the person wearing it.

Ondova develops an intelligent sensing platform, licensed through SDKs and reference designs. It lets your device recognize its wearer, follow heart rate, read wear state and taps, and supply the context that personalized audio and wellness features depend on.

How it works Design partners
No added sensorsThe speaker you already ship
Works during playbackSensing continues while music plays
LicensedSDK and reference designs
In-ear, on-ear, over-earOne front end, three form factors
The problem

Billions of devices already sit on the best continuous sensing site on the body, and almost none of them understand the person wearing them.

Adding that understanding today means adding dedicated sensors, and with them cost, power, board space and integration effort that most products cannot absorb. Wearer-awareness stays locked inside a handful of flagship products.

Ondova removes that barrier, starting at the ear: in-ear, on-ear and over-ear, from earbuds and headphones to communication headsets and hearing devices.

How it works

The speaker at your ear becomes the sensor.

Heart rate, touch and in-ear detection usually need their own parts. Here they come from the speaker already there.

No added sensors. Device makers license Ondova software and the design of a small circuit, and get new features from a part their product already has.

Every earbud and headphone has a speaker at the ear. The way that speaker responds changes slightly with the ear around it and with tiny movements, such as a heartbeat or a tap. Ondova reads those changes, even while music plays.

Close view of an ear wearing a graphite earbud. Marks show the earbud's speaker, a heartbeat at the entrance of the sealed ear canal and a tap on the earbud. This speaker is the sensorIt plays sound and feels the ear A tapon the earbud A heartbeatin the sealed ear canal
Illustrative photo. Ondova technology goes into device makers' own earbuds and headphones.

  1. The speaker feels the ear

    How it fits in the ear, a heartbeat, a tap: each one slightly changes how the speaker responds.

  2. A small circuit reads it

    It sits on the speaker's wires and separates that change from the music playing.

  3. Software turns it into features

    • Wearer identificationEach ear canal responds a little differently.
    • Heart rate
    • Wear detection
    • Tap and gesture
    • Fit and seal In validation
The signal path

The speaker is already a precise transducer coupled to the ear.

A driver does not only push air. It responds to the ear it sits in, and to the body around it: its own electrical behaviour carries the load of the ear canal and the small movements of the tissue around it. Ondova reads that response through a compact analog front end placed in the driver path, so the speaker becomes the sensing element. Playback is cancelled in the analog domain, which is why sensing keeps running while music plays.

The result is a speaker-sensed, driver-derived signal carried over the audio interfaces your device already has. Nothing optical, capacitive or inertial is added to the bill of materials.

This adds a capability your device does not have today, and it works alongside the sensors you already ship.

  1. 1Driver in the earoutput device and sensing element
  2. 2Analog front endin the driver path, playback cancelled in the analog domain
  3. Standard audio interface
  4. 3Ondova SDKwearer identity, heart rate, wear state, taps
  5. Nothing optical, capacitive or inertial is added.
No added sensorsThe sensing element is the driver you already specify.
No dedicated sensor to powerThe sensing element is a component the device already drives. The front end adds a small analog block.
Housing and tuning stay yoursThe front end sits in the electrical driver path. Its effect on the driver load is documented in the reference design.
Capabilities

What the signal supports.

Every capability below comes from the same speaker-sensed path. New features arrive as software on the same front end, each one validated before we call it a capability.

01

Wearer identification

The device recognizes an enrolled wearer from the signature of that person's ear canal. Personalization, profiles and correct device assignment on shared or rotating headsets. Enrolment is opt-in and the wearer's signature stays within the product.

02

Heart rate

Heart rate read from the driver signal while music plays. Intended for general wellness and evaluated against a chest-strap reference.

03

Wear detection

Worn or removed, for auto-pause, auto-resume and power saving, without a dedicated presence sensor.

04

Tap and gesture

Button and touch behaviour with no capacitive touch hardware, useful where sealing, cost or industrial design rules a touch surface out.

05
In validation

Fit and seal

A fit test from the same signal, so the wearer can be told when the seal is wrong and the tuning is not being delivered.

Conceptual illustration of a speaker driver with rings of response spreading outward, some marked and some still open

Beyond the validated set

In validation
Acoustic-profile personalization Voice activity detection Respiratory rate
Research
Heart-rate variability trends Leakage-aware EQ assist

All wellness features are intended for general wellness. Ondova makes no medical or clinical claims. Performance figures are shared with evaluating partners under NDA, together with the test conditions behind them.

Integration

Three ways the platform reaches your product.

What changesOne small analog block per driver path, one spare ADC input on the codec to digitize the sensed signal, and the Ondova estimation layer where your product runs its software.
What staysThe driver, the housing and the tuning. The amplifier and playback chain. The sensors you already ship. The radio and its qualification.
SDK

The estimation layer

It takes the speaker-sensed stream and returns wearer identity, heart rate, wear state and taps through a documented API per capability.

  • Host implementation for PC and phone, used for evaluation and wired products
  • Embedded targets discussed with partners under NDA
  • Output as events and streams, so the product decides what it exposes
Reference design

The analog front end

A documented per-side block for the driver path.

  • Schematic and component guidance
  • Documented effect on the driver load and on the amplifier
  • Characterized once per headphone model during bring-up
Evaluation kit

Hardware to try it on

Reference hardware fitted with the Ondova front end, and a demo application, so your engineers can run the validated capabilities themselves.

  • Live demo of wearer identification, heart rate, wear detection and tap
  • Your own form factor where practical, characterized with you during the evaluation
  • Findings report against targets we define together, with the protocol attached

The technology inside your product.

Ondova is an ingredient brand: we license the sensing intelligence to the companies that build the devices, and the product and the customer relationship stay yours. The analog front end ships as a reference design today and is built to be integrated into the hardware you already choose, including partner silicon.

Powered by Ondova
Evidence

A skeptical engineer should be able to reproduce it.

Our commercial case rests on the signal being real, so we treat validation as part of the product. Evaluating partners receive the protocol with every result, and we state the limits of our datasets. We welcome independent co-validation.

Request the validation pack
ReferenceGround truth per capability: a chest-strap reference for heart rate, labelled sessions for wearer identity, wear events and taps.
DevicesForm factor, wired or wireless, and the reference-design revision used.
PopulationParticipant count, with the limits of the set stated.
ScenariosSeated and quiet, office ambient, music playing, walking and cycling, plus the scenarios your use case depends on.
WindowsThe averaging window stated alongside every metric.
ReportingOne page per evaluation: conditions, metrics, plots and limitations.
Design partners

Evaluate it on your own terms.

A structured evaluation, typically six to eight weeks, on reference hardware fitted with the Ondova front end or, where practical, on your own form factor. We run a small number of evaluations per cycle so that each partner works directly with the engineers who built the signal path.

  • An evaluation kit and a demo application covering the validated capabilities
  • Weekly engineering syncs with our core team
  • A findings report measured against targets we define together

What we ask for: a named technical owner, test units where applicable, and a mutual NDA.

Start a conversation Ask for the partner overview

Who this fits

Makers of earbuds, headphones, communication headsets and hearing devices with an active product cycle. Operators running shared or rotating headsets at scale, where knowing who is wearing a device, and whether it is being worn, changes how the operation runs.

Company

Built in Montreal.

Ondova Technologies Inc. builds the sensing layer for ear-worn devices. Founded in Montreal through TandemLaunch, a deep-tech venture studio.

Patrick Huang
Patrick HuangCo-founder and Chief Executive Officer. Strategy, partnerships, OEM engagement. San Diego.
Danilo Pena
Danilo PenaCo-founder and Chief Technology Officer. Signal path, algorithms, IP. Montreal.

Working with us

Open roles at Ondova →
Vision
Ondova becomes the intelligent sensing platform inside the next generation of hearables and edge AI devices.

Today the work is the ear, where the installed base is largest and our validation is furthest along. If a device can be made to understand its wearer through hardware it already has, that is Ondova's work.

Ondova does not make consumer products and makes no medical claims. Our work is the sensing layer, delivered through the partners who build the devices and the silicon.