Instrumental Assessment of Voice Flashcards
1
Q
Instrumental Measurements of Voice: Clinical Utility
A
DETECTION
– identify the existence of a voice problem
SEVERITY
– assess the severity or stage of progression of the voice
DIAGNOSIS
– identify the differential source of the voice problem
2
Q
Instrumental Measures in the Voice Laboratory
A
- Acoustic recording and analysis
- aerodynamicassessment
- laryngealvideostroboscopy
- high speed digital imaging (HSDI)
- electroglottography
- electromyography
3
Q
Acoustic Analyses
A
- Assess physical correlates of perceptual evaluation
- Pitch/Frequency
- Loudness/ Intensity
- Perturbation
- Noise
- Spectral/cepstral analyses
4
Q
Acoustic Recording and Analysis
A
- Provides an imperfect, non-invasive measure of vocal function
- Can discriminate normal from pathologic voice
- Is inconsistent agreement between acoustic measures and audio-perceptual
ratings of voice quality - Can measure change in vocal production across time
5
Q
=Acoustic Analyses
A
- Fundamental Frequency
- Intensity
- Frequency range
- Jitter/shimmer/cycle-to-cycle variability
- Noise to harmonics ratio
- Phonetogram/Voice Range Profile (frequency by intensity plots)
- Cepstral Analysis (CPP)
6
Q
Routine Voice Acoustic Measures
Fundamental Frequency (F0) (1. level)
A
- Directly reflects the vibration rate of the vocal folds; the acoustic correlate of pitch
- Unit of measurement: Hertz (Hz) or cycles per second
- Normative data = 100-150 Hz males; 180-250 Hz females
- Know male and female F0!
- May be measured from sustained vowels, reading, or conversation
- Useful to estimate the appropriateness of F0 for sex and age and for demonstrating
pre- and post-treatment change
7
Q
Frequency Variability (2. variability)
A
- Pitch sigma is the standard deviation of the fundamental frequency
- Assess and documents variation of F0 during speech production
- How much is the pitch changing during task or during recording in general
- Should not have a lot of variability during vowel prolongation
- Should have more variability during reading and conversational speech
- How much voice in varying in normal sample
8
Q
Phonation range (3. range)
A
- Range of frequencies from the highest to the lowest that a patient can produce
- How much client can physically do: see what system is capable of (increase and decrease pitch as high or as low as they can with modal/falsetto pitch)
- May be expressed in Hz or semitones
- Semitones: 12 semitones make up 1 octave
- Normal young adults have about a 3-octave range; may vary with practice
- Octave = doubling of frequency
- 1 octave: 100-200 Hz, 2
octave: 200-400 Hz, etc.
9
Q
Frequency Perturbation (4. perturbation) JITTER
A
- The change of frequency from one successive period to the next (horizontal)
- Assessing cycle to cycle variability in the voice: are cycles similar to each other
- More similar = more periodic = better quality
- Unit of measurement is jitter; several algorithms are used to extract jitter
- Normative data (normal voice) for jitter percent is less than 1.00%
- Hoarseness = will be more than 1%
- Measures MUST be made from sustained vowels ONLY: not reading or speech
- Reading and speech has frequency variability which will increase jitter score - May represent variation of vocal fold mass, tension, muscle activity, or neural activity
all of which may affect the periodicity of vocal fold vibration
10
Q
Intensity (I0) (1. level)
A
- Directly reflects the sound pressure level (SPL) of voice; direct correlate of loudness
- Reflects vocal fold adduction
- Unit of measure is the logarithmic decibel (dB) scale
- May be measured from sustained vowels, reading, or conversation
- Useful as pre- and post-treatment measure
- Overall SPL:
- Average SPL in dB (depends on
room you are getting measures
from)
- Indication of the strength of
vocal fold vibration (Norms: 75-80
dB conversation)
11
Q
Amplitude Variability (2. variability)
A
- Standard deviation of the SPL during connected speech
- Reflects loudness variability
- Reflects phase closure
- Sustained vowels: want variability to be minimal
- Reading and speech: will vary more
12
Q
Dynamic range (3. range)
A
- Range of vocal intensities that a person can produce (Norms: 50-115 dB SPL)
- General range that you expect them to be in, don’t have to go to 50 and 115,
need to be within the average range (which was not specified) - Checking the loudest they can go and softest that they can go
13
Q
Amplitude Perturbation (4. perturbation) SHIMMER
A
- Small cycle to cycle changes of the amplitude of the vocal fold signal (vertical)
- Unit of measurement is shimmer; several algorithms are used to extract shimmer
- Normative data for shimmer dB is less than .35 dB
- Measures MUST be determined from sustained vowels ONLY: not reading and
speech - May represent variation of vocal fold mass, tension, muscle activity, or neural activity
all of which may affect the amplitude of vocal fold vibration
14
Q
Harmonics to Noise Ratio
A
- A ratio measure of the energy in the voice signal over the noise energy in the voice signal (signal they are getting should be greater than noise in the room OR the amount of “noise” occurring from vocal productions)
- Noise: breathiness, raspiness, hoarseness, roughness
- May be derived from different algorithms and expressed in various units
- Greater signal or harmonic energy in the voice reflects better voice quality
- Large noise energy represents more abnormal function
- Ideally greater than 20 dB
- Most reliable in comparison to jitter and shimmer
- Jitter and shimmer: can only get data from sustained vowel and software has
difficulty marking the beginning and end of cycles - More severe hoarseness = less reliable jitter and shimmer values
14
Q
Voice Range Profile (Phonetogram)
A
- Plots maximum and minimum intensities (loudest/softest) for entire frequency range
- Assess F0 and intensity and individual’s minimum and maximum capabilities.
- Resulting plot is ellipsoid-shaped frequency/intensity profile and the dimensions are expressed in semitones
- Most useful in pre and post-treatment of professional voice users because it provides a thorough description of the patient’s physiologic limits of frequency and intensity
- More therapy = range increases = holes/gaps should fill, ellipsoid should increase
15
Q
Spectral Analyses
A
- Includes spectrogram, line spectrum, cepstrum
- Useful to assess the interaction between the glottal sound source (vocal folds) and supraglottic (vocal tract) influences
16
Q
Spectral Analyses: - Spectrogram
A
- Displays the glottal sound source and filtering characteristics across time
- Putting many photographs together (like a video)
- Both formant frequency energy (vocal tract resonance) and noise components
(aperiodicity) are presented in a 3D scale - Horizontal axis = time — vertical axis = frequency
- Lowest band = F0; formants lie above
- Gray scale (darkness) represents intensity change
- Dark gray bands = stronger formant energy
- Light gray bands =poor formant energies and nosier signal
17
Q
Spectral Analyses: Line
A
- Fourier transformation of spectrogram (snapshot in time, like a photograph)
- Plot in the frequency domain
- Plots all harmonics/frequencies at a single time point
spectrum- Frequency on the horizontal axis, amplitude on the vertical axis
18
Q
Spectral Analyses: Cepstral analyses
A
- Fourier transformation of the spectrum
- Emphasize peaks of the strongest harmonics, including fundamental frequency
- Cepstral peak prominence (CPP) compares the height of the cepstral peak to
the aperiodic noise in the voice - Sharper the peak, more periodic the voice (breathy → less prominent peak)
- Cepstral measures do not rely on pitch detection like perturbation measures
- Measures not dependent on identifying fundamental frequency, marking cycle
boundaries or adequate intensity - Cepstral measures have strong correlations with perceptual CAPE-V measures
- Measures of cepstral analyses
- Cepstral peak prominence (CPP) in dB
- CPPvowel: average of 3 trials: 1-second sample of steady /a/
- CPPspeech: using Rainbow Passage or CAPE-V sentences
- Normatives vary by software: Praat, SpeechTool, ADSV
