Quantifying the Voice: Clinical Perceptual Scaling and Visi-Pitch Acoustic Analysis

Foundations of Voice Quantification: The Clinician vs. The Machine

• The Clinician's Ear: Auditory-Perceptual Evaluation

  • Nature: This approach is qualitative and listener-dependent.

  • Purpose: It aims to capture the holistic human experience of voice quality by identifying specific characteristics through the clinician's expertise.

  • Core Characteristics Tracked:

    • Roughness.

    • Breathiness.

    • Strain.

  • Primary Tools: Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V) and the GRBAS scale.

• The Diagnostic Machine: Instrumental Acoustic Analysis

  • Nature: This approach is quantitative and driven by biofeedback.

  • Purpose: Acoustic data is utilized to validate and quantify clinical perception by measuring the exact physical properties of sound waves.

  • Core Properties Measured:

    • Frequency.

    • Amplitude.

    • Perturbation.

  • Primary Tools: Visi-Pitch (specifically modules like Real-Time Pitch and Multidimensional Voice Program [MDVP]).

Standardized Perceptual Scales: GRBAS and CAPE-V

• The GRBAS Scale

  • Protocol: Evaluation is conducted using samples of spontaneous speech.

  • Scale Type: Equal-appearing interval scale.

  • Scoring Range: $0$ to $3$ ($0 = \text{Normal}$, $3 = \text{Severe}$).

  • Metrics Measured (The GRBAS Acronym):

    • G (Grade): The overall severity of the voice abnormality.

    • R (Roughness): Perception of irregular vocal fold vibration.

    • B (Breathiness): Perception of air leakage through the glottis.

    • A (Asthenia): Perceived weakness or lack of power in the voice.

    • S (Strain): Perception of excessive vocal effort or hyperfunction.

• The CAPE-V Scale

  • Protocol: Evaluation is conducted using sustained phonation (vowels) and specific sentences.

  • Scale Type: $100\text{\textbackslash,mm}$ Visual Analog Scale (VAS).

  • Scoring Range: $0$ to $100$ ($0 = \text{Normal}$, $100 = \text{Severely Deviant}$).

  • Metrics Measured:

    • Overall Severity.

    • Roughness.

    • Breathiness.

    • Strain.

    • Pitch.

    • Loudness.

  • Key Insight on Asthenia: Asthenia is explicitly excluded from the CAPE-V. It was determined to be highly influenced by loudness and difficult for clinicians to parse apart perceptually from other factors.

• CAPE-V Cutoff Boundaries

  • Clinical categories are established by translating visual analog lines into distinct segments on a $100\text{\textbackslash,mm}$ ruler.

  • Overall Severity / Grade Mapping:

    • Normal (Consistent with GRBAS 0): $0-15\text{\textbackslash,mm}$.

    • Mild (Consistent with GRBAS 1): $16-39\text{\textbackslash,mm}$.

    • Moderate (Consistent with GRBAS 2): $40-69\text{\textbackslash,mm}$.

    • Severe (Consistent with GRBAS 3): $70-100\text{\textbackslash,mm}$.

  • Metric-Specific Boundaries:

    • Roughness: Mild ($14-37\text{\textbackslash,mm}$), Moderate ($38-66\text{\textbackslash,mm}$), Severe ($66-100\text{\textbackslash,mm}$).

    • Breathiness: Mild ($15-34\text{\textbackslash,mm}$), Moderate ($35-68\text{\textbackslash,mm}$), Severe ($69-100\text{\textbackslash,mm}$).

    • Strain: Defined primarily in the Normal range ($0-15\text{\textbackslash,mm}$).

Visi-Pitch: Standardized Software Overview

• Definition: An online, standardized software program (often part of the Sona-Speech system) designed to assess disordered voice and speech through the use of visual biofeedback.

• Primary Clinical Applications:

  • Routine therapy tasks.

  • Evaluating progress (e.g., tracking changes in pitch range over time).

  • Visual cueing for patients.

  • Determining voice typing.

  • Specialized use cases: Motor speech disorders, fluency, auditory rehabilitation, and accent modification.

• Critical Diagnostic Limitation: Visi-Pitch is intended for assessment and performance tracking. It is not a standalone diagnostic tool. While results provide objective biofeedback regarding current performance, they do not determine the underlying medical etiology or disorder.

The Visi-Pitch Ecosystem: Two Analytical Lenses

• Module 1: Real-Time Pitch (The Macro View)

  • Focus: Evaluation of prosody, vocal range, and connected speech.

  • Feedback: Provides real-time visual data on pitch, energy, and time parameters.

  • Key Clinical Tasks Captured:

    • Habitual Pitch ($F_0$)

      • talking out of habitual pitch can lead to nodules

    • Maximum Phonation Time (MPT).

      • phonatory function, which is crucial for evaluating vocal health and efficiency in patients. A shortened MPT may indicate vocal strain or pathology.

    • Monotone Evaluation.

    • Pitch Range.

    • Loudness.

• Module 2: Multi-Dimensional Voice Program (MDVP) (The Micro View)

  • Focus: Analysis of vocal fold perturbation and noise values.

  • Input: Derived from a single sustained vocalization (typically the vowel /a/).

  • Key Clinical Metrics Captured:

    • Jitter.

    • Shimmer.

    • Relative Average Perturbation (RAP).

    • Noise-to-Harmonic Ratio (NHR).

    • Voice Turbulence Index (VTI).

Real-Time Pitch: The Five Macro Tasks and Protocol

• Detailed Task Breakdown:

  1. Habitual Pitch ($F_0$): Represents the average pitch during connected speech. It serves as the natural baseline for the patient.

  2. Maximum Phonation Time (MPT): Measured as the longest sustained /a/ produced on a single breath. This is an indicator of respiratory-phonatory efficiency.

  3. Monotone Evaluation: Assessment of prosody while the patient reads standardized passages (e.g., "The Grandfather Passage" or "The Rainbow Passage"). Flat prosody may indicate neurological or affective issues.

  4. Pitch Range: Evaluated via vocal glides from the patient's lowest to highest possible pitch. A restricted range can indicate pathology or the effects of aging.

  5. Loudness ($dB$): Measures vocal intensity in $dB\text{\textbackslash,SPL}$. This assesses if the patient is using appropriate and consistent energy levels during phonation.

• Operational UI Process Flow:

  1. Launch: Open Sona-Speech and select the 'Real-Time Pitch' module.

  2. Execute: Follow on-screen instructions for specific tasks (e.g., reading or gliding). Navigate to the 'Energy Tab' specifically for loudness measurements.

  3. Capture: Click 'Stop' and then 'Compute Result Statistics'.

  4. Document: Capture a clear photo of the results screen. Clear the page using the 'X' icon (top right) to prepare for the next trial.

Acoustic Biomarkers: Real-Time Pitch Norms

• Habitual Pitch (Average $F_0$):

  • Female Baseline: $180-250\text{\textbackslash,Hz}$ (Average: $\text{\textbackslash,approx 199-220\text{\textbackslash,Hz}}$).

  • Male Baseline: $100-150\text{\textbackslash,Hz}$ (Average: $\text{\textbackslash,approx 109-125\text{\textbackslash,Hz}}$).

• Maximum Phonation Time (MPT):

  • General Adult Norm: $15-25\text{\textbackslash,seconds}$.

  • Female Target: $15\text{\textbackslash,seconds}$.

  • Male Target: $20\text{\textbackslash,seconds}$.

• Pitch Range:

  • Female Range: $165\text{\textbackslash,Hz} - 275\text{\textbackslash,Hz}$ (Musical equivalent: Semitones $E_3 - G_4$).

  • Male Range: $85\text{\textbackslash,Hz} - 185\text{\textbackslash,Hz}$ (Musical equivalent: Semitones $E_2 - G_3$).

  • Clinical Expectation: Approximately $2-3$ octaves ($24-36\text{\textbackslash,semitones}$).

• Loudness:

  • Conversational Average: $\text{\textbackslash,approx 60-70\text{\textbackslash,dB}}$.

  • Typical Range: $46-98\text{\textbackslash,dB}$.

MDVP Perturbation Matrix and Analysis Protocol

• The Golden Rule of MDVP: Analysis requires a Green Signal Only. The patient must phonate a steady, comfortable /a/. If the recording signal enters the "red zone," the extraction of perturbation data will be invalid.

• Analysis Execution Steps:

  1. Navigate to Sona-Speech and select "MDVP".

  2. Click 'File' then 'Record' (Shortcut: $F_{12}$).

  3. Patient produces a steady /a/ while the clinician ensures a green signal is maintained.

  4. Click 'Stop'.

  5. Navigate to 'Protocol' and select "Complete MDVP Analysis". Take a photo.

  6. Navigate to 'Protocol' and select "Show MDVP Parameters in Active Window" (Shortcut: $F_8$). Take a photo.

  7. Go to 'Window' and select "Purge Active Window" to clear the current data.

• MDVP Metric Definitions and Thresholds:

  • Jitter (%): Measures pitch instability between cycles. Clinically sounds like pitch breaks or roughness. Threshold: < 1.0\text{\textbackslash,\text{\textbackslash,}\%} (Ideally $0.5\text{\textbackslash,\text{\textbackslash,}\%}$).

    • pitch breaks

    • vocal folds are not openeing and closing in a steady manner

    • Should be moving in a sequence, steady manner

  • Shimmer (%): Measures amplitude instability/variability between cycles. Perceived clinically as hoarseness or roughness. Threshold: < 3.5\text{\textbackslash,\text{\textbackslash,}\%} (General cutoff < 5\text{\textbackslash,\text{\textbackslash,}\%}).

    • amplitude (sometimes high, sometimes low)

  • Relative Average Perturbation (RAP): Represents smoothed pitch instability. Clinically indicates irregular vocal fold vibration. Threshold: < 0.5\text{\textbackslash,\text{\textbackslash,}\%} (Average observed: $0.68\text{\textbackslash,\text{\textbackslash,}\%}$).

  • Noise-to-Harmonic Ratio (NHR): Compares noise components against periodic sound. Perceived as a breathy, rough quality. Threshold: < 0.19.

  • Voice Turbulence Index (VTI): Measures high-frequency turbulence. Indicates weak or breathy phonation typical of loose vocal fold adduction. Threshold: < 0.02 (Average observed: $0.061$).

Synthesis: Correlating Perceptual and Acoustic Findings

• Acoustic Validation of Clinical Perception:

  • Subjective Finding: Severe Roughness on the CAPE-V.

    • Objective Correlate: Elevated Shimmer and Jitter on MDVP (reflecting irregular vibration and amplitude variations).

  • Subjective Finding: Severe Breathiness on the CAPE-V.

    • Objective Correlate: Elevated NHR and VTI on MDVP (reflecting high noise/turbulence from incomplete glottal adduction).

  • Subjective Finding: Monotone conversational speech.

    • Objective Correlate: Restricted Pitch Range and flat $F_0$ variance on Real-Time Pitch modules.

• Final Summary Principle: The Visi-Pitch system does not replace the expertise of the clinician's ear; it serves to provide a quantifiable, objective dimension to the clinician's perceptual findings.