Frequently Asked Questions

No. PhonaLab measures physical properties of the voice signal — fundamental frequency, perturbation, cepstral peak prominence, etc. — which are language-independent. The validated indices (AVQI, ABI, CPPS, CSID) were originally normed on specific language populations (Dutch, American English, Brazilian Portuguese), so clinical cutoff values may vary across languages. However, the acoustic measurements themselves are universally applicable.

No. PhonaLab uses deterministic signal processing algorithms, not machine learning models. The acoustic parameters are computed using established methods from the voice science literature — the same algorithms used in Praat, implemented via Parselmouth. This means: • No training data is required • Results are reproducible and deterministic • The same audio file will always produce the same measurements

AI (Claude by Anthropic) is used as a development tool to help write code, design the user interface, and draft documentation. AI is not involved in the acoustic analysis itself. The voice measurements come entirely from signal processing algorithms (Parselmouth/Praat). AI does not process, interpret, or influence the acoustic results.

PhonaLab uses Parselmouth, which embeds Praat's acoustic algorithms (version 6.1.38). The measurements are algorithmically identical to Praat. A validation manuscript comparing PhonaLab's AVQI, ABI, CPPS, and CSID against desktop Praat using the PVQD database (290 speakers) has been submitted to the Journal of Voice.

Voice Analyzer: F0 (mean, SD, range), jitter (local, RAP, PPQ5), shimmer (local, APQ3, APQ5, APQ11), HNR, CPPS AVQI & ABI Calculator: AVQI, ABI, CPP, HNR, shimmer local, shimmer local dB, slope, tilt Spectral Analysis: LTAS, Alpha Ratio, CSID, L1-L0, CPP Pitch Visualizer: F0 contour, F1-F3 formants MPT & DSI Calculator: Maximum Phonation Time, Dysphonia Severity Index

Green indicates values within normal range, yellow suggests borderline values that warrant monitoring, and red indicates values outside the typical range that may require clinical attention. These thresholds are based on published normative data.

For reliable acoustic analysis, follow these guidelines: Environment: • Record in a quiet room with minimal background noise • Avoid rooms with hard surfaces that create echo • Turn off fans, air conditioning, and other noise sources Phone position: • Hold the phone at a consistent distance of 10–15 cm (4–6 inches) from the mouth • Position at a 45° angle to the side of the mouth to reduce breath noise • Keep the phone steady during recording Recording settings: • Use the highest quality audio setting available • Ensure sample rate is at least 22,050 Hz (most phones default to 44,100 Hz, which is fine) • Disable noise cancellation or audio enhancement features if possible Voice task: • For sustained vowel (/a/), produce a steady sound at comfortable pitch and loudness • Sustain for at least 3–4 seconds • Aim for stable, consistent phonation — avoid starting or trailing off

Minimum for AVQI/ABI: 22,050 Hz (spectral energy in the 6–10 kHz range is required) Recommended: 44,100 Hz (standard for most recording devices) Important: Do not downsample to 16 kHz — this removes high-frequency information needed for multiparametric indices

PhonaLab accepts WAV, MP3, and M4A files. WAV (uncompressed) is preferred for highest accuracy, but compressed formats typically produce comparable results for clinical purposes.

For professional assessments, a headset microphone or external USB microphone positioned 4–10 cm from the mouth provides the most reliable measurements. Built-in laptop or smartphone microphones can work for screening purposes.

No. PhonaLab uses a stateless architecture. Audio files are processed in memory and immediately discarded after analysis. No voice recordings are stored on our servers. This design: • Protects patient privacy by default • Simplifies regulatory compliance • Means users must retain their own recordings if needed

PhonaLab is designed to be "HIPAA aware" — it minimizes data handling by not storing any patient audio or identifiable information. However, full HIPAA compliance depends on the clinician's overall workflow. Clinicians should: • Not include patient names or identifiers in uploaded filenames • Store recordings and reports in their own HIPAA-compliant systems • Follow their institution's data governance policies

We use industry-standard encryption for all data transmission. Account information is stored securely with Supabase, which employs enterprise-grade security measures including encryption at rest and in transit.

Yes. PhonaLab is freely available for research use. For collaboration inquiries, contact phonalab@gmail.com.

A validation manuscript is currently under review at the Journal of Voice. Once published, a formal citation will be provided. In the meantime: PhonaLab (https://www.phonalab.com). Acoustic voice analysis platform developed by J. C. Lucero, University of Brasília.

Yes. All acoustic analysis algorithms are based on established, peer-reviewed methods: • AVQI follows Maryn et al. • ABI follows Maryn et al. (Brazilian Portuguese validation) • CSID follows Awan & Roy (2012) • Core acoustic measurements use Praat algorithms via Parselmouth

PhonaLab works on modern browsers including Chrome, Firefox, Safari, and Edge. Chrome is recommended for best microphone recording support.

Yes. PhonaLab is fully responsive and works on smartphones and tablets. The browser microphone recording feature works well on mobile. For file uploads, desktop browsers may be more convenient.

Account creation enables: • Analysis history tracking (for your reference) • Consistent access across devices • Future features like saved reports and session comparisons Registration uses passwordless magic link authentication — no password required.

Analysis may fail if: • The recording is too short (less than 1 second) • The audio is too quiet or contains excessive background noise • The sample rate is below 22,050 Hz • The file format is not supported Try re-recording in a quieter environment with adequate volume.