Hear everything again

NEXT-GENERATION
SOUND PERSONALIZATION

Full-spectrum hearing correction for music, on headphones and on speakers, calibrated to the device and measured at the ear.

In software, for any product that plays sound. Available to integrate.

HEAR IT THROUGH THEIR EARS

This interactive demo lets you hear Altitude's sound personalization as it ships today in ARIA ONE. Press play, then use the switch to flip between the original song and the personalized version. Choose one of three listeners to hear the personalization made for them. Turn on "Their ears" to hear the song the way that listener hears it every day, and then what it sounds like once it has been personalized for them.

loading audio
Listener
Ears
0:00 / 0:00
Sound
Now playing

Both versions play at the same level. Differences you hear are the personalization.

About this listener

Alex · the average listener

Alex hears the way most adults do. The highest sounds have faded a little, the kind of change most people never notice. Voices lose some of their clarity and the shimmer at the top of the music sits further back.

Based on the average result of listeners who have taken the Altitude hearing test.

ONE SYSTEM, FOUR MODULES

Altitude's sound personalization is a complete hearing system for music, film, and speech across four licensable modules. The modules may be licensed together as one system with a single integration, or individually, depending on the product you are building.

Hearing test

Full spectrum, 125 Hz to 20 kHz, up to 37 frequencies per ear, calibrated to the device.

Hearing correction

An EQ and a limiter, fitted together to the measured ear. Lossless, full spectrum, built for music. Stereo modes for headphones and for speakers.

Crossfeed

Speaker-like spatialization on headphones. Simulates the acoustic experience of listening to speakers in a room. Reduces listening fatigue and enhances spatial awareness.

Corrective EQ

A final correction stage for imported frequency-response curves. Headphone correction for listeners, or behind-the-scenes voicing and batch-based adjustment by engineers.

ARCHITECTURE

Distribute sound personalization to fit your product. Hearing test and DSP controls in the app, yours or Altitude's. Hearing-profile fitting and preset sync in the backend service. DSP compute in your SoC, or in your software app when the product is the app.

ALTITUDE CLOUD Fitting · AI fitting assistant profile + preset sync · licensing test results up fitted profile + presets down APP · PHONE OR DESKTOP YOURS OR ALTITUDE'S Hearing test · calibrated per device Control surface · 6 views · reskin or rebuild profile + presets, local coefficients BLE · network · or in-process no cloud link in the audio path YOUR PRODUCT · AUDIO PATH earbuds speakers car TV app audio · decoded ALTITUDE · IN YOUR SoC OR APP hearing EQ → limiter crossfeed · corrective EQ (optional) brand tuning · user EQ DAC · amp · transducer → listener ANC / transparency the product's own hardware path · not touched

Headphones and earbuds

Correction in the SoC, so every source is processed. Or in the app for a pilot without firmware work.

Speakers and soundbars

Correction in the DSP of speakers, soundbars, streamers, car audio, and other device-side systems, with modes built for open-air listening.

Software apps and services

Test, fitting, and DSP in the app that makes the sound: music streaming, video, media playback, voice, and calls.

Hearing-care apps

Test and fitting in the app, correction applied to streamed media as a listening layer beside the hearing aid.

THE HEARING TEST

Personalization can only be as accurate as the measurement it starts from.

FullScale measures hearing from 125 Hz to 20 kHz at up to 37 frequencies per ear, on the listener's own headphones or earbuds, calibrated per device.

Validated at the eardrum: across 1,582 tones on three reference devices, 125 Hz to 20 kHz, 99% were delivered within the tolerance ANSI S3.6 sets for clinical audiometers, with a mean offset of -0.5 dB.

Delivered vs forecast

Error by frequency, dB at the eardrum

2026 mean2024 mean±1 sdtarget range
±3 dB±5 dB-12-8-40+4+8+12125 Hz250 Hz500 Hz1 kHz2 kHz4 kHz8 kHz16 kHz
Error distribution

How far each tone landed from its forecast

Count of tones per 0.5 dB · purple = inside ±3 dB · beyond ±10 dB collected in the end bars · 1,582 tones
-10 dB-5 dB-3 dB0 dB+3 dB+5 dB+10 dB

FullScale is not an audiometer and makes no medical claim. The ANSI S3.6 tolerance is used only as a reference for accuracy.

Top 3 reference devices, 2026: Technics EAH-AZ80, Apple AirPods Pro (2nd gen), KEF Mu3. 2024: Technics EAH-AZ60, Apple AirPods Pro (2nd gen), KEF Mu3.

Explore every measured tone →

Full validation results
Top three reference devices, 500 Hz to 8 kHz
100% of 915 tones within tolerance, mean offset -0.20 dB
Top three reference devices, 125 Hz to 16 kHz
100% of 1,522 tones, mean offset -0.36 dB
Top three reference devices, 125 Hz to 20 kHz
99% of 1,582 tones, mean offset -0.47 dB
Reference devices
2026: Technics EAH-AZ80, Apple AirPods Pro (2nd gen), KEF Mu3. 2024: Technics EAH-AZ60, Apple AirPods Pro (2nd gen), KEF Mu3
Tolerance
±3 dB to 5 kHz, ±5 dB above, the accuracy ANSI S3.6 specifies for audiometers
Method
One placement per device, no adjustment, GRAS 45CC-17 headphone test fixture, Audio Precision APx517B analyzer. Tones at the microphone noise floor and frequencies a device cannot reach are not counted. Measured October 2024 and 27 October 2026

FullScale is not an audiometer and makes no medical claim. The ANSI S3.6 tolerance is used only as a reference for accuracy.

THE HEARING CORRECTION

Personalization is only as effective as the DSP behind it.

Altitude's hearing correction carries an accurate hearing profile into sound personalization for music, film and speech. Full-spectrum correction for a full-spectrum profile: fitted per ear from the measured thresholds, as a high-resolution filter and dynamics stage working across the complete audible range.

Three stereo modes cover headphone listening and open-air speaker listening.

Using the same performance predictors used for hearing health devices, the system shows measurable gains in speech intelligibility, music quality, and audibility.

HASPI v2 · speech in noise 0.32 → 0.69 standard moderate loss (N3)
HASPI v2 · speech in noise 0.64 → 0.96 our 60+ population average; equal to a normal ear on the same output
HASPI v2 · speech in noise 0.88 → 0.95 our whole-population average
HAAQI · music quality 0.46 → 0.54 standard moderate loss (N3), released master
HAAQI · music quality 0.68 → 0.87 our 60+ population average, released master
HAAQI · music quality 0.88 → 0.92 our whole-population average, released master
Speech Intelligibility Index against playback level for the N3 listener, unprocessed and corrected
Predicted intelligibility (SII) for the N3 listener against playback level. The corrected master at 80 dB sits above the unprocessed curve at every level.
Full results, three hearing profiles
Profiles
Three hearing profiles were used in this analysis: the N3 standard moderate hearing-loss profile from Bisgaard (2010); the average profile of Altitude listeners aged 60 and over; and the average profile of all listeners in Altitude's hearing-test database. Together, they show how the system performs for a standard reference case, an older-listener case, and a general-listener case.
HASPI v2, speech intelligibility in noise
HASPI v2 is a published predictor of how intelligible speech is for a listener in background noise. In this analysis, speech was presented 3.7 dB above the noise bed at 80 dB SPL. The hearing correction raised the score from 0.32 to 0.69 for the N3 profile, from 0.64 to 0.96 for the 60+ average profile, and from 0.88 to 0.95 for the whole-population average. For the 60+ profile, the corrected score reached the same 0.96 value as a normal-hearing listener at the same output.
SII, music audibility
SII is a published measure of how audible important parts of a signal are to a listener. Here it was used on a released music master at 80 dB SPL to estimate how much of the program remains audible for each hearing profile. The hearing correction raised the score from 0.33 to 0.42 for the N3 profile and from 0.44 to 0.47 for the 60+ average profile. For the whole-population average, the score did not change, which suggests that audibility in this case was already limited more by masking in the music than by hearing loss. Across the tested playback range, the unprocessed master did not reach the corrected value.
HAAQI, music quality
HAAQI is a published predictor of perceived music quality for listeners with hearing loss. In this analysis, the hearing correction raised the score from 0.46 to 0.54 for the N3 profile, from 0.68 to 0.87 for the 60+ average profile, and from 0.88 to 0.92 for the whole-population average. The largest improvement appeared in the 60+ average profile, while the whole-population average also improved.
HASQI v2, speech quality
HASQI v2 is a published predictor of perceived speech quality for listeners with hearing loss. In this analysis, it was used to compare the processed speech signal against clean speech for the N3 hearing profile. The hearing correction raised the score from 0.28 to 0.46, indicating that more of the original speech quality was preserved for that listener.
SII, speech in noise
SII is a published measure of how much of a speech signal is audible to a listener above their hearing thresholds and the background noise. Here it was applied to speech presented 3.7 dB above a noise bed at 80 dB SPL. For the N3 profile, the hearing correction raised the score from 0.40 to 0.55 as delivered, and from 0.40 to 0.48 when the output was level-matched to the input. That suggests part of the gain comes from the shape of the correction and part from the level it adds.
Vocal-to-instrumental ratio, music
This measure looks at how clearly the singer stands out from the band in music. After processing, the vocal and instrumental were separated and compared across the speech bands. The hearing correction raised the vocal-to-instrumental ratio by 1.30 dB for the N3 profile, 1.16 dB for the 60+ average profile, and 0.76 dB for the whole-population average. For the 60+ profile, the quietest vocal details gained 3 to 13 dB against the band between 4 and 8 kHz. This is a measured change in prominence, not a listening-test result.
Vocal image
This measure looks at how tightly the singer sits in the centre of the stereo image. The vocal and the band were separated after processing and measured independently. The full chain raised the vocal's centre-image score from 0.77 to 0.88 for the N3 and 60+ profiles, and to 0.84 for the whole-population average, while the band stayed close to its original width. The result is a voice pulled more clearly into the centre without narrowing the rest of the music. This is a measured image property, not a listening-test result.
What the metrics are
HASPI v2 and HAAQI are published predictors used by HearAdvisor to score hearing aids. HAAQI is also the music-quality metric used in the Cadenza challenge, and SII is defined by ANSI S3.5.

THE DYNAMICS STAGE

Sound personalization adds gain to music that is already at the ceiling. Altitude's personalization includes an advanced dynamics stage that delivers the correction at full strength without being turned down. Against FabFilter Pro-L 2, the $199 reference limiter in professional audio, on the same pre-master mix at the same drive and the same ceiling, it produced more loudness.

Loudness · same drive, same ceiling +0.76 LU integrated, BS.1770-4, against Pro-L 2 on the same mix
Perceived loudness +4.4 % sones, ISO 532-1, loudest passage
Loudness range · at equal loudness +0.31 LU 2.72 vs 2.41 LU, with more dynamics kept at the same level
Short-term loudness over the whole track for the source mix, the Pro-L 2 master and the Altitude master
Short-term loudness over the whole track. Same mix, same 8.4 dB drive, same -0.2 dB ceiling. The two limited masters track each other with a near-constant offset.

CROSSFEED AND CORRECTIVE EQ

Sound personalization is the core of the system, but it is not the whole offering. Crossfeed brings speaker-like spatialization to headphones, and Corrective EQ applies imported correction curves for headphone correction, room correction, and device voicing. Both may be licensed alongside personalization or on their own.

Crossfeed

A headphone-listening processor that makes stereo playback feel more speaker-like. It reduces the in-your-head effect of headphones while preserving width, focus, and detail.

Crossfeed 2 →

Corrective EQ

A correction stage for imported response curves. It applies headphone correction, room correction, and device voicing from measurements you already trust, with higher filter quality than a simple band EQ.

Corrective EQ →

HOW TO INTEGRATE

Where the hearing test resides dictates the shape of integration. That choice determines where the controls live, what gets reused, and how much of the existing product surface moves into your environment.

The three pathways below show the practical integration shapes: use Altitude's app, attach your own controls to the system, or package the full experience inside your product.

Your hardware, our app

Where the test lives
ARIA ONE, calibrated and branded for your device
Where the controls live
ARIA ONE
Where the DSP runs
Your firmware, or the app
Altitude provides
DSP port, device calibration, branding, and hosting
You provide
Hardware samples, a firmware contact, and approvals
Time
Shortest path; reuses the existing app and keeps the work focused on the DSP port.

Each pathway uses the same underlying DSP and fitting system. What changes is where the test lives, where the controls live, and how much of the existing product surface is reused.

A WORKING CONTROL SURFACE

Altitude already ships working control surfaces for sound personalization and the surrounding DSP modules in ARIA ONE on iOS and Android. Supporting flows for personalization fitting, profile selection, and preset management already exist within the navigation stack. Together, these form the core UI a partner would need to implement if hosting the controls inside the product app. The existing surface provides a clear starting point that can be reskinned, simplified, or rebuilt for the product.

Controls

Supporting views

ARIA ONE mobile sound home: active profile, signal path, current response and module cardsARIA ONE Hearing Essentials card: three moods, drive, correction, loud preset with soft to hard sliderARIA ONE Advanced Hearing card: stereo mode buttons, smooth, intensity, MS width and loud togglesARIA ONE Crossfeed card: three presets, mix, and an advanced group of five parametersARIA ONE presets and profiles browser: full, limiter, EQ and crossfeed preset lists, factory and user, with save, resave and deleteARIA ONE AI assistant: ask, preset, help and learn modes with a message field

HOW THE COMMERCIAL MODEL WORKS

Evaluation is free. Shipping is licensed. Integration support is available if you want me involved.

A partner can evaluate the current system at no cost. Shipping it in a product requires a production license. Some teams may integrate on their own; others may want me to port the DSP, package the hearing test, attach the control surface, and support launch. Ongoing support and hosted services are covered through the commercial agreement.

Free to evaluate

Evaluation

Evaluate the current system at no cost to confirm product fit, the likely integration path, and whether it belongs in your product. If you want evaluation on your target SoC or DSP framework, you provide the development kit and any required framework access or licenses.

Required to ship

Production license

If you want to ship the technology in a product, you license it for that use. The production license covers the right to use the system in the agreed product and commercial context. Pricing is defined per product and scope.

Optional help · ongoing support

Integration and support

You can integrate the system yourself, or I can be involved directly. Paid support can cover the DSP port, calibration work, hearing-test packaging, control-surface attachment, launch support, updates, hosted fitting services, new device calibrations, and agreed support terms.

Evaluation
Free on the current system
On-target evaluation
Partner supplies development kit and any required framework access
Production use
Licensed to ship
Integration help
Optional, paid, and scoped to the work required
Ongoing support
Can cover updates, hosted services, new calibrations, and agreed support terms
Pricing
Defined per product and scope
IP
Patent pending, PCT/US2024/036841, national phase. Nothing transfers during evaluation.

Tell me what you want to build, and which chip or software platform you plan to use. If the fit is there, I can build toward your product from the start.

COMMON QUESTIONS

Does this work with Spotify Connect, AirPlay, and Chromecast?

Yes, when the DSP runs inside the device. Casting sends audio straight to the headphone or speaker, so device-side integration corrects every source. App-side integration corrects what plays through that app.

Can it fit our chipset, DSP budget, and latency target?

Usually, yes. The DSP can be shaped to different compute budgets, from embedded SoCs to app processors, and latency depends on the chosen processing configuration. Share the part, signal path, and constraints and I can tell you what fits.

Who owns the hearing data?

The listener owns the hearing data. Storage and sync can live on the device, on your infrastructure, or on ours, depending on the integration shape.

Can this live inside our app?

Yes. The test and controls can stay in Altitude's app, be opened from your app, or be packaged inside your app. The right starting point depends on timeline, scope, and how much of the experience you want to own.

Can we start with one module or one integration path and expand later?

Yes. The modules can be licensed together or individually, and teams can start with the fastest integration path before moving deeper. A product can begin with one module or one control path, then expand once the fit is proven.

Is this a medical device?

No. This is sound personalization for music and media, not a medical device or hearing aid. It measures hearing to fit playback, but it does not diagnose, treat, or make medical claims.

Does this replace our house sound?

No. The system works out of the box, but it is also designed to be voiced for the brand. If you want a different sonic presentation, it can be adjusted through presets or tuned with Altitude to match the sound you want to achieve.

START THE CONVERSATION

Tell me what you want to build, and which chip or software platform you plan to use. If the fit is there, I can build toward your product from the start.