phonation measurement 1

Question 1

whats the diff bw jitter and shimmer?

Answer 1

• jitter: stability of F0 across time (closely related to pitch)
• shimmer: stability of intensity across time (closely related to loudness)

Question 2

whats the diff bw HNR and VOT?

Answer 2

• HNR (harmonic : noise ratio): overall spectral characteristics of sound
• VOT (voice onset time): timing of vocal fold activity relative to other speech events

Question 3

what are 3 problems with evaluating F0 based on pitch?

Answer 3

1. pitch judgements are influenced by top-down info
2. pitch judgements are non-linear
3. pitch is not a pure index of F0

Question 4

T or F: humans are more sensitive to pitch variations at high frequencies than low frequencies

Answer 4

false – more sensitive to pitch variations at low frequencies

Question 5

how many semitones are in an octave?

Answer 5

12

Question 6

what is the relationship between going up an octave and frequency?

Answer 6

each octave up = doubling of frequency (in other words: each octave down = halving of frequency)

Question 7

the amount of F0 change is different at lower vs higher frequencies. the ear is more sensitive at the ____ end.

Answer 7

lower

Question 8

what is peak-picking? (3)

Answer 8

• Old analog method that detects peaks in signal above a threshold
• Creates a pulse for each peak
• Counts the pulses per second to get F0

Question 9

Explain what counting zero-crossings is (2)

Answer 9

• Counts every time signal crosses 0 (x-axis)
• Need to double your half cycle to get correct values (??)

Question 10

Why are peak-picking and zero-crossings not ideal for speech samples? Is there any solutions?

Answer 10

• Saw tooth signal = highly irregular – difficult to set threshold, inaccurate data
• Note: some improvement is possible via low-pass filtering

Question 11

What is accelerometry? What are its pros and cons?

Answer 11

• Accelerometry: uses accelerometers which are sensitive to body vibrations and can detect vocal fold movements
• Pros: cheap, easy, and yields clean signal
• Cons: have to build it yourself

Question 12

How can you get F0 from a wide band spectrogram?

Answer 12

Count vertical striations within a time frame: striations / time frame (convert time to seconds)

Question 13

How can you get F0 from a narrow band spectrogram?

Answer 13

Look at first harmonic (horizontal striations)

Question 14

T or F: vowel based F0 is different from speaking F0

Answer 14

True

Question 15

Why do we need to use a big chunk of passage sample (i.e., 14 seconds) to measure pitch? (2)

Answer 15

• Because vowels vs consonants change pitch… i.e., pitch is higher for /i/ than /a/
• Accuracy of +/- 3Hz requires ~14 seconds

Question 16

T or F: using the same sentence instead of similar but slightly different phonetic tokens reduces errors in SF0 estimates

Answer 16

True

Question 17

T or F: variability of single sentences and entire passages correlate high

Answer 17

false

Question 18

What is the male vs female average range for SF0?

Answer 18

• Males: 90-180Hz
• Females: 180-225Hz

Question 19

Are SF0 values for children w/ Down syndrome different than typical children?

Answer 19

Not reliably

Question 20

Are SF0 values for people w/ voice disorders uniformly lower?

Answer 20

No – except for cases of VF edema

Question 21

Do people who stutter tend to have smaller SF0 ranges?

Answer 21

Yes

Question 22

Anything that increases VF thickness affects VF vibration by…

Answer 22

Slowing vibrations

Question 23

What is a maximum phonational frequency range (MPFR)?

Answer 23

Index of physical condition of phonatory system

Question 24

Is MPFR influenced by age or physical condition?

Answer 24

Physical condition

Question 25

How do we find MPFR?

Answer 25

Phonetogram

Question 26

A phonetogram maps out the relationship between F0 and ____ over the individual’s vocal range.

Answer 26

Vocal intensity (dB)

Question 27

What might a phonetogram look like for a person with a voice disorder relative to a person without a voice disorder?

Answer 27

Reduced in both frequency and intensity

Question 28

Define jitter (2)

Answer 28

• Frequency perturbation
• Index of short-term time stability of phonatory changes in frequency

Question 29

What are jitter measurements mostly limited to?

Answer 29

Sustained vowel productions

Question 30

Define shimmer

Answer 30

Cycle-to-cycle variations in amplitude

Question 31

What is one way to calculate jitter? (4)

Answer 31

1. Take period duration of each cycle and subtract that from next one
2. Store absolute values
3. SUM the differences between the successive cycles
4. Divide that SUM by the number of cycles you processed

Question 32

What is the relationship bw jitter and mean F0? How can we correct this bias?

Answer 32

• Jitter tends to show inverse relationship with mean F0
• Divide mean perturbation by mean period duration (jitter ratio)

Question 33

What should the jitter number be?

Answer 33

Less than 1% (i.e., if phonating at 100Hz, jitter should be 99-101)

Question 34

Is there more jitter at the ends of a signal or middle?

Answer 34

More at ends (that’s why we measure midportion of vowels)

Question 35

T or F: jitter ratio decreases with vocal intensity

Answer 35

True

Question 36

Which vowels may have lower jitter?

Answer 36

/i/ (bc tenses vocal folds)

Question 37

Who tends to have higher jitter values, males or females?

Answer 37

females

Question 38

Is jitter correlated with perceived hoarseness?

Answer 38

Yes – if higher than 1%