Voice Assessment & Instrumentation SPE 682 Fall 2024 PDF

Voice Assessment & Instrumentation SPE 682 Fall 2024 Teresa Kiernan, MA, CCC-SLP General Assessment Principles 1. What are we measuring? 2. Reliable, Repeatable, & Valid 3. Use sets of measures 4. Normative data when possible 5. Standardized scores when possible 6. Ma...

Voice Assessment & Instrumentation SPE 682 Fall 2024 Teresa Kiernan, MA, CCC-SLP General Assessment Principles 1. What are we measuring? 2. Reliable, Repeatable, & Valid 3. Use sets of measures 4. Normative data when possible 5. Standardized scores when possible 6. Make results accessible and shareable to the patient 7. Connect the results to the patient Man against Machine!  Utility of Instrumental Measures Can the instrument: 1. Detect a problem 2. Assess the severity 3. ID the source of the issue 4. Serve as a tool for education or biofeedback Pros and Cons of Objective Measures PROS CONS 1. Insight into how the voice 1. Can’t tell us what we hear system is behaving 2. Only as good as the user 2. Documents vocal status in 3. Costly graphical, objective manner 4. Numerous and not always 3. Easy to compare pre- and comparable post- 5. Needs to be reliable 4. Can compare to norms (repeatable) and valid 5. Can be used as visual (measure what it is feedback or to explain the supposed to measure) problem in a concrete way 7 Limitations of Perceptual Voice Assessment Measures the product of the system; not how subsystems coordinate. ○ Can’t isolate the subsystems (e.g., breath from sound source) Wide inter-rater variability Somewhat subjective Can’t compare to normative measures Can’t judge small, progressive changes over time  Clinical Utility 1. Does the measure reflect the voice in singing or speech? 2. Is the measure reliable across trials? 3. Are there norms to compare your data to? 4. Is the protocol efficient? What are we measuring? Acoustic Signal INDIRECT Aerodynamic changes measures in pressure and flow Visual Images We make inferences about laryngeal pathology from the results What Equipment do you need? Microphone Digital (ideal: Head recording Mounted) PRE-AMP! equipment Keeps consistent distance from mouth Internal or external Angled from the center sound card of the mouth Sample at least 20KHz Cardioid or 16 Bits of amplitude unidirectional quantization Avoid clipping Quiet recording space 11 Adapted from Awan et al., 2015  Headsets Logitech PC Headset 120 Shure BETA 53 Acoustic Measurements of Vocal Function Measurement Types ○ Spectral acoustic analysis (time-based) ○ Cepstral acoustic analysis (frequency-based) Acoustic Characteristics Spectral Cepstral ○ Fundamental frequency (F0) ○ Intensity (dB) ○ “noise” and perturbation to periodicity Spectral Methods: Frequency & Intensity Frequency (rate of vocal fold vibrations – periods / second) Fundamental Frequency (f0) ○ Lowest frequency (Hz) produced by oscillation of the vocal folds (perceptual: pitch) Harmonics: ○ component frequencies of oscillations of wave ○ Sometimes referred to as “overtones” Intensity (Amplitude of signal)  Spectral Methods: Sound Perturbations (Voice Quality Measurements) Waveform ○ Jitter: cycle-to-cycle differences in frequency ○ Shimmer: cycle-to-cycle differences in intensity (amplitude) ○ Noise-to-harmonic ratio (NHR) Spectrogram ○ Type (visualized) ○ Noise-to-harmonic ratio (NHR) Pros & Cons of Spectral Methods Pros Cons ∙ Output very sensitive to variables: ∙ More objective than auditory-perceptual o Dysphonic voices: Does not work as well assessment (e.g., less influenced by clinician for dysphonic voices bias & order effect) o Pitch, voicing, & intonation: Only on /a/, not ∙ Easy to acquire with appropriate equipment as reliable for connected speech ∙ Correlations between spectral parameters, and the severity, quality, and type of pathology have been shown to be moderate at best Cepstral Acoustic Analysis In contrast to traditional acoustic analyses, cepstral analysis can extract f0, harmonics, amplitude, and noise-to-harmonics without depending on cyclical (time-based) boundaries. Spectral Cepstral  Cepstral Peak Prominence (CPP) In a typical voice, the voice signal (f0, harmonic structure) will be more “prominent” than noise in the signal. Higher peaks in normal phonation and lower in breathy phonation, compared to the line (regression). Measured in dB, although can also extract f0 (Hz) and harmonics The Cepstral/Spectral Index of Dysphonia (CSID) The acoustic correlate to the CAPE-V auditory-perceptual VAS ratings. Uses algorithm to capture voice parameters (pitch/frequency, loudness/intensity, and quality) in multivariate manner using the “we were away a year ago” sentence and sustained vowel /a/. Doesn’t take into account listener experiences or articulatory characteristics of the speaker. Normed on relatively small sample size (n=160). Pros & Cons of Cepstral Methods Pros Cons ∙ Can be used with both ∙ Can still get inaccurate sustained and measurements if there is lots of spontaneous/running background noise, inadequate speech signal-to-mic distance, etc. ∙ Can more accurately (garbage in is still garbage out!) measure dysphonic voices  Voice & Speech Tasks Running Speech Cape-V sentences Spontaneous speech: How did you get here today? Standardized passage: The rainbow passage Sustained vowel Comfortable pitch (3-5 seconds) X 3 High pitch (3-5 seconds) X 3 Low pitch (3-5 seconds) X 3 Soft as possible X 3 Loud as possible X 3 26 Awan et al., 2015 Acoustic Analysis Equipment Analysis software ○ Use of standard (documented) algorithms ○ In wide clinical use ○ Formally validated via comparison with other commonly used programs Common software ○ Pentax Medical computerized speech lab (Multi-Dimensional Voice Program; MDVP) ○ Praat Acoustic Analysis Apps Advanced Spectrum Analyzer Pro (F0 & dB) Voice Pitch Analyzer (F0) Sound Analyzer (dB) Sonneta Voice Monitor (F0, dB, MPT) Voice Test (F0, Shimmer & Jitter %) Voice Analyst (F0, dB, MPT) Pitche Analyzer (F0 & dB) NIOSH SLM (dB) AERODYNAMIC ASSESSMENT Aerodynamic Analysis Quantitative measures of laryngeal function in relation to pulmonary and articulatory systems Assessment Parameters: ○ Airflow ○ Air pressure ○ Airway resistance Aerodynamic Assessment Parameters Airflow:Estimates amount of air needed to sustain vocal fold oscillations using supraglottic airflow measured during vowel. Airpressure: Subglottal pressure required to initiate vocal fold oscillation. Estimated by intraoral air pressure w/ voiceless stop consonant /p/. Airwayresistance: Product of laryngeal airflow and subglottal pressure. Quantifies the aerodynamic power needed to initiate and sustain vocal fold oscillation for a given glottal configuration.  Pressure Measurement Methods Phonatory Aerodynamic System (PAS) Subglottal pressure with speech (Indirect) PAS in Action PAS in Action 2  Low-Cost Solutions Maximal Phonation Time ○ Not without it’s problems!! Must do repeat trials to get average S/Z Ratio ○ Requires repetition of each sound 2-3x ○ Norm = 0.7 – 1.0 Phonation Quotient = MPT / Vital Capacity (w/spirometer) Windmill Spirometer Digital Spirometer Putting it all together… Indirect measures of acoustics and aerodynamics should be cross-referenced with your perceptual and visual voice and speech behaviors.