Measurement that meets your standards.
Validated parameters, established surveys, and a longitudinal patient view that runs entirely on the patient's device. Built for SLPs, ENT clinics, and voice-therapy programmes.
- Range
- A2 — F♯5
- On target
- 86%
- Stability
- Good
- Take
- 0:08
Numbers your colleagues already trust.
Voxavia's DSP is calibrated against gold-standard references. Parameter outputs are framed in published norms — and the validation study against Praat / MDVP is on the roadmap.
| Parameter | What it captures | Reference |
|---|---|---|
| F0 | Fundamental frequency from a sustained vowel. | Praat / MDVP |
| Jitter | Cycle-to-cycle frequency perturbation. | MDVP |
| Shimmer | Cycle-to-cycle amplitude perturbation. | MDVP |
| HNR | Harmonics-to-noise ratio. | Praat |
| CPPS | Cepstral peak prominence smoothed — robust voice-quality scalar. | Hillenbrand 1996 |
| F1 / F2 / F3 | First three formants via LPC, every ~200 ms. | Praat |
| GNE | Glottal-to-noise excitation ratio. | Michaelis 1997 |
| Subharmonic ratio | Diplophonia / period-doubling detector. | Sun 2002 |
| LTAS slope | Long-term average spectrum slope. | Mendoza 1996 |
| MFCC | Mel-frequency cepstral coefficients. | Davis & Mermelstein 1980 |
- F0Praat / MDVP
Fundamental frequency from a sustained vowel.
- JitterMDVP
Cycle-to-cycle frequency perturbation.
- ShimmerMDVP
Cycle-to-cycle amplitude perturbation.
- HNRPraat
Harmonics-to-noise ratio.
- CPPSHillenbrand 1996
Cepstral peak prominence smoothed — robust voice-quality scalar.
- F1 / F2 / F3Praat
First three formants via LPC, every ~200 ms.
- GNEMichaelis 1997
Glottal-to-noise excitation ratio.
- Subharmonic ratioSun 2002
Diplophonia / period-doubling detector.
- LTAS slopeMendoza 1996
Long-term average spectrum slope.
- MFCCDavis & Mermelstein 1980
Mel-frequency cepstral coefficients.
Self-awareness, not diagnosis.
Voxavia presents established voice and reflux surveys to patients as self-rating tools. Their results show up in the patient's progress chart, not as a clinical conclusion.
Voice Handicap Index
Patient-reported impact across physical, functional, and emotional sub-scales.
Reference: Jacobson 1997
Grade · Roughness · Breathiness · Asthenia · Strain
Hirano's perceptual self-rating, captured as patient-reported.
Reference: Hirano 1981
Reflux Symptom Index
9-item index for laryngopharyngeal reflux symptoms.
Reference: Belafsky 2002
No PHI ever leaves the device.
Voxavia has no backend. Patient audio, surveys, and longitudinal metrics live in IndexedDB on the patient's machine — never on a server. Patients can export and bring sessions to their appointment as a portable JSON / ZIP.
- 01
On-device IndexedDB
Sessions, audio blobs, and PBs stay in the browser.
- 02
No analytics SDK
Zero third-party trackers in either the app or the marketing site.
- 03
No account required
Patients can use the app immediately. Reduces friction at intake.
- 04
COPPA-aware kid surface
When the kid tier ships, no social features and parental-consent flows are non-negotiable.
Clinical privacy posture
As of 2026
- Patient audio
- Local IndexedDB
- Patient identifiers
- None collected
- HIPAA / GDPR exposure
- None — no transit
- Server-side storage
- None
- Cookies
- Theme preference only
Validation roadmap.
Clinicians don't trust unvalidated numbers, so we don't ask you to. Here's the candid status.
Calibration against Praat / MDVP
DSP parameters are calibrated against published norms today. Per-parameter calibration documents are in development.
Validation study
A peer-reviewed comparison against Praat / MDVP on a multi-speaker corpus. Targeted submission once we hit the right sample size.
Practitioner mode
Therapist creates a plan, sends a share link, sees only the patient's chosen metrics. Strong B2B angle, on the roadmap.
A trial, on your terms.
We pilot with clinics that want measurement embedded in their workflow. Tell us what you're treating and we'll set you up.