Lab 1: Acoustics

1
Lab 1 Acoustics

• 8.30-8.35/12.30-12.35 Introduction, expectations
• 8.35-9.45/1235-145 (9.00/1.00 hand out
  assignment)
• In pairs Basic Acoustics Spectrograms Module
  of Sensimetrics, SECTIONS 1-4, 6
• 9.45-10.10/1.45-210
• Group Discuss Research and Clinical Application
  of Acoustic Analysis CHALLENGE QUESTIONS
• a. Voices of Singers example
• b. Esophageal Speech example
• c. Parkinsons-Depression example
• 10.10-10.20/2.10-2.20 Questions, more
  Sensimetrics time, time for assignment

2
REMINDER!!

• The research studies that we discuss in lab are
  meant to be a challenge, to expand your knowledge
  about how the things you are learning in 425 lab
  relate to a variety of topics in speech and
  hearing sciences and disorders.
• Discussions are meant to expose you to a wider
  breadth of areas of current research in applying
  acoustic analysis techniques in clinical
  research, without focusing on the details or
  methodologies involved.
• These discussions are also an exercise for you to
  become more fluent in obtaining information
  directly from peer-reviewed papers in the field.
• While you are required to think critically about
  these discussions, you will not be formally
  tested on content.

3
Effects of Singing Training on the Speaking Voice
of Voice Majors(Mendes et al 2004, J of Voice)

• This longitudinal study gathered data with regard
  to the question Does singing training have an
  effect on the speaking voice? 14 voice majors
  (12F, 2M, 17-20) were recorded once a semester
  for four consecutive semesters, while sustaining
  vowels and reading the Rainbow Passage.
  Acoustic measures included speaking fundamental
  frequency (SFF) and sound pressure level (SPL).
  Perturbation measures included jitter, shimmer,
  and harmonic-to-noise ratio. Temporal measures
  included sentence, consonant, and dipthong
  durations.

4
Abstract, Continued

• Results revealed that, as the number of semesters
  increased, the SFF increased while jitter and
  shimmer slightly decreased. Repeated measure
  analysis, however, indicated that none of the
  acoustic, temporal, or perturbation differences
  were statistically significant. These results
  confirm earlier cross-sectional studies that
  compared singers with non-singers, in that
  singing training mostly affects the singing voice
  and rarely the speaking voice.

5
Singing Exp, Challenge Question

• Methodological Consideration
• Q You know how difficult it can be to obtain
  exact measurements from spectrograms.
• In order to avoid error measurements in small
  temporal measures, such as closure duration in
  /t/ and dipthong durations, what could be done?

• A Intra-measurer agreements from 2 people
• (r0.76-0.88)

6
Acoustic Analysis of Tracheo-Oesophageal versus
Oesophageal Speech (Debruyne et al 1994, J. of
Laryngology and Otology)

• BACKGROUND
• These are 2 common options for voice rehab after
  total laryngectomy
• Both use the pharyngo-oesophageal segment, which
  consists of mucosa and supporting tissues in the
  area surrounding the pharynx to the oesohagus
• T-OE uses pulmonary expiratory air (from lungs)
• OE uses ejected air from the oesophagus

7
Tracheo-oesophageal Speech Abstract

• In order to evaluate the vocal quality of
  tracheo-oesophageal and oesophageal speech,
  several objective acoustic parameters were
  measured in the acoustic waveform (F0, jitter,
  shimmer) and in the spectrum (harmonic
  prominence, spectral slope). 12 patients using
  tracheo-oesophageal speech (with the Provox
  valve) and 12 patients using oesophageal speech
  for at least 2 months, participated.

8
Abstract, continued

• The main results were that Tracheo-oesophageal
  speech results in greater loudness and longer
  phonatory duration. Tracheo-oesophageal voices
  more often showed a detectable F0, and that this
  F0 was fairly stable there was also a tendancy
  to more clearly defined harmonics and less
  perturbation in tracheo-oesophageal speech. This
  suggests a more regular vibratory pattern in the
  pharyngo-oesophageal segment.
• So, a better quality of the voice can be
  expected, in addition to the longer phonation
  time and higher maximal intensity, in T-OE
  speech.

9
T-OE vs OE speech Challenge Question

• Q What might be a reason for better overall
  vocal quality with tracheo-oesophageal speech
  compared to oesophageal speech?

Hint
How do those two methods differ in terms of
speech production?
A Air flow from the lungs is a more efficient
driving force compared to the short ejection of
air out of the oesophagus. BUT, you cant
ignore clinical considerations Higher pressure
is needed to initiate and sustain vibration in
the T-OE method than in regular vocal fold
vibration.
10
Parkinsons or Depression?(Flint, et al., 1992,
J of Psycholinguistic Research)

• In its early stages, Parkinson's disease (P.D.)
  may be difficult to distinguish from major
  depression (M.D.) leading to inappropriate
  management. Both illnesses are characterized by
  psychomotor retardation. The neurovegetative
  symptoms used to diagnose M.D. are not specific
  and in P.D. may be due to the physical illness
  itself. Currently, differentiation of the two
  disorders relies on subjective clinical
  observation. Improved diagnostic accuracy based
  on more objective data is needed.

11
Abstract, Continued

• To this end, this study used computerized
  acoustic analysis to contrast speech patterns in
  P.D. and M.D. The sample consisted of 30 P.D.
  patients without depression or dementia, 30
  patients with uncomplicated M.D., and 31 normal
  controls, each 60 years of age or over. Of the
  acoustic variables studied, M.D. patients had
  significantly reduced rates of speech compared
  with P.D. patients. The data suggest that this
  temporal measure of speech may be useful in the
  differentiation of P.D. and M.D.

12
Parkinsons vs. Depression CHALLENGE QUESTION

• Q Since this paper says that speech rate can be
  used as a diagnostic tool, cant clinicians
  simply measure speech rate with a tape recorder
  and a stopwatch? Why or why not?

A Sure, it can be measured that way, but that
might not be a sensitive enough measure.
Measuring rate with full acoustic analysis tools
can allow clinicians to measure not just how many
syllables per second, but also how long each
individual segment (such as a vowel) is over
time. It can also allow clinicians to measure
the gaps of time between segments. It is these
more precise measurements that can be used as
diagnostic tools. So, rate of speech can mean
a lot more than just words or syllables per
minute or second.
13
Lab 2 Vowels

• 8.30-9.45/12.30-1.45
• In pairs Speech Acoustics Vowel Acoustics
  Module of Sensimetrics, SECTIONS 1-5
• (9.15/1.15 Hand Out Assignment)
• 9.45-10.10/1.45-2.10
• Group Discuss Research and Clinical Application
  of Acoustic Vowel Analysis CHALLENGE QUESTIONS
• a. Maxillectomy example
• b. Children with hearing impairment example
• 10.10-10.20/2.10-2.20
• Questions on Lab 1 or Lab 2

14
Digital Acoustic Analysis of Five Vowels in
Maxillectomy Patients(Sumita, et al 2002, J. of
Oral Rehab)
BACKGROUND What is Maxillectomy?

• Full or partial removal of the maxilla (upper
  jaw) and hard palate.
• Used in cases of oral tumors (carcinoma)
• May or may not be fitted with prostheses
  (artificial replacement)

15
Digital Acoustic Analysis of Five Vowels in
Maxillectomy Patients(Sumita, et al 2002, J. of
Oral Rehab)

• The aim of the study was to characterize the
  acoustics of vowel articulation in maxillectomy
  patients. Digital acoustic analysis of five
  vowels, /a/, /e/, /i/, /o/, and /u/, was
  performed on 12 male maxillectomy patients, and
  12 normal male individuals.
• A simple set of acoustic descriptions called the
  first and second formant frequencies, F1 and F2,
  were employed and calculated based on linear
  predictive coding.

16
Abstract, Continued

• The maxillectomy patients had a significantly
  lower F2 for all 5 vowels and a significantly
  higher F1 for only the /i/ vowel. From the data
  plotted on an F1-F2 plane in each subject, we
  determined the F1 range and F2 range, which are
  the differences between the minimum and the
  maximum frequencies among the 5 vowels.
• The maxillectomy patients had a significantly
  narrower F2 range than the normal controls. In
  contrast, there was no significant difference in
  the F1 range.

17
Abstract, Continued

• These results suggest that the maxillectomy
  patients had difficulty controlling F2 properly.
  In addition, the speech intelligibility (SI) test
  was performed to verify the results of this new
  frequency range method. A high correlation
  between the F2 range and the score of the SI test
  was demonstrated, suggesting that the F2 range is
  effective in evaluating the speech ability of
  maxillectomy patients.

18
Maxillectomy Patients, CHALLENGE QUESTION

• Q Why would ONLY /i/, and not any other vowel,
  show a higher F1 in the patients?

/i/ is a high vowel. Where is F1 normally?
HINT 1 HINT 2 HINT 3

• Picture the Vowel Space! (F1 corresponds
  with what? F2 corresponds with what?)

What part of the oral cavity is most physically
affected?
19

• A A relatively lower position of the tongue
  because of a defect in the front upper palate
  increases the F1 value. This increase is not
  apparent for the other high vowel /u/ because /u/
  is a back vowel.
• The other vowels (/a/, /e/, /o/) would also not
  be affected, since they are low (ie, the tongue
  position is low relative to the palate).

20
An Acoustic Metric of Assessing Change in Vowel
Production by Profoundly Hearing-Impaired
Children(Fourakis, et al, 1993 JASA)

• The purpose of this study was to investigate the
  feasibility of developing an acoustic metric to
  assess vowel production in profoundly HI
  children. The approach taken was to develop a
  metric from acoustic analysis of vowel
  productions and then compare it with the
  perceptual ratings of the same productions by
  listeners. Speech samples were collected from 3
  profoundly HI children.

21
Abstract, continued

• The metric used extracted the fundamental and
  first, second, and third formant frequencies to
  represent the tokens as points in a 3-dimensional
  auditory-perceptual space modeled after earlier
  work by other researchers. Euclidean distances
  were determined between each point and the
  intended vowel, which was represented by
  coordinates taken from other research. The data
  suggest that this 3D metric provides significant
  correlations between production and perception.

22
Vowels of Children with HI CHALLENGE QUESTION

• Q The main purpose of this research is to
  develop and evaluate an acoustic measure of the
  goodness of vowel productions by children with
  HI. What do you think are other factors, besides
  vowel triangle sizes, that may affect goodness
  ratings?

Think about how some people with profound HI
might sound during speech production
HINT
23

• A Vowel production by children (and adults) with
  HI tend to be extremely long (gt400ms), can
  contain combinations of steady-state and glide
  components (ie dipthongized), or can be very
  nasalized.
• Measures such as these should ideally be
  incorporated into such a metric, to that
  acquisition of correct phonetic targets can be
  successfully tracked.

24
Lab 3 Consonants

• 8.30-9.45/12.30-1.45
• In pairs Speech Acoustics Consonant
  Acoustics Module of Sensimetrics, ALL SECTIONS
• (9.15 Hand Out Assignment)
• 9.45-10.10/1.45-2.10
• Group Discuss Research and Clinical Application
  of Acoustic Consonant Analysis
• a. VOT example
• b. Motor Speech Disorders example
• 10.10-10.20/2.10-2.20
• Questions

25
CHALLENGE QUESTION comes first this time!

• Q Based on what you know about acoustic
  analysis, how could you measure VOT or initial
  stop consonants, like p, t, k?

• A Waveform Spectrogram using VF vibration
  onset,
• start of F0, or start of formants (5 common ways)

• Q Do you think one method is better or easier
  than another, or do you think it doesnt make a
  difference which you use? Why?

• A We shall see which method proves best!

26
Accuracy and Variability of Acoustic Measures of
Voicing Onset(Francis, et al, 2003, JASA)

• Five commonly used methods for determining the
  onset of voicing of syllable-initial stop
  consonants were compared. The speech and glottal
  activity of 16 native speakers of Cantonese with
  normal voice quality were investigated during the
  production of CV syllables in Cantonese.
  Syllables consisted of the initial consonants
  /ph/, /th/, /kh/, /p/, /t/, and /k/ followed by
  the vowel /a/. All syllables had a high level
  tone, and were all real words in Cantonese.

27
Abstract, continued

• Measurements of voicing onset were made based on
  the onset of periodicity in the acoustic
  waveform, and on spectrographic measures of the
  onset of a voicing bar (F0), the onset of the
  first formant (F1), second formant (F2), and
  third formant (F3). These measures were then
  compared against the onset of the glottal opening
  as determined by electroglottography (gold
  standard). Both accuracy and variability of
  each measure were calculated.

28
Abstract, continued

• Results suggest that the presence of aspiration
  in a syllable decreased the accuracy and
  increased the variability of spectrogram-based
  measurements, but did not strongly affect
  measurements made from the acoustic waveform.
  Overall, the acoustic waveform provided the most
  accurate estimate of voicing onset measurements
  made from the amplitude waveform were also the
  least variable of the five measures. These
  results can be explained as a consequence of
  differences in spectral tilt of the voicing
  source in breathy versus modal phonation.

29
Acoustic Analysis in Motor Speech Disorders
BACKGROUND

• Acoustic analysis of dysarthric speech is
  challenging because the dysarthrias can be
  complex disorders with potential disruptions
  occurring throughout the speech production
  system. Some disruptions may mask others, and
  the acoustic signal can be greatly diminished in
  the contrasts that are needed for precise
  measurements.
• Acoustic analysis can be informative because it
  affords quantitative analyses that carry
  potential for subsystem description and for
  determining the correlates of perceptual
  judgements of intelligibility, quality, and
  dysarthria type.

30
Toward an Acoustic Typology of Motor Speech
Disorders(Kent et al, 2003, Clinical Linguistics
Phonetics)

• Acoustic methods have progressed to the point
  that an acoustic typology of the motor speech
  disorders can be construed from a parametric
  assessment of the speech subsystems (e.g.,
  phonation, nasal resonance, vowel articulation,
  consonant articulation, intonation, and rhythm).
  The results of this analysis can be interpreted
  in respect to global functions (e.g., voice
  quality, intelligibility, and prosody).

31
Abstract, Continued

• This paper reviews studies showing that specific
  acoustic analyses have demonstrated potential
  value toward the overall goal of constructing
  acoustic profiles of dysarthria and apraxia of
  speech. Several different acoustic measures are
  relevant to the study of motor speech disorders,
  and these are increasingly supported by normative
  data and by guidelines for clinical application.
• Examples of these applications are discussed for
  a variety of specific neurologic diseases or
  perceptual types of disorders. Acoustic studies
  are useful in the study of motor speech disorders
  and recent progress points to a parametric
  analysis.

32
Some Specific Examples

• 1. Especially when the tongue is paralytic
  (cannot move), hypokinetic (moves too much), or
  atrophic (decreased mass), the area of the F1-F2
  quadrilateral can be an index of the degree of
  lingual impairment, on the assumption that
  restricted mobility of the tongue is reflected by
  abnormalities in F1-F2 patterns. Compression of
  the acoustic vowel space is a general property of
  many dysarthrias that is related to
  intelligibility.

33

• 2. Measures of formant slope typically are
  obtained from analysis of words containing
  conspicuous formant transitions associated with
  phonetic transitions (e.g., consonant to vowel).
  Studies show a consistent relationship between
  speech intelligibility and average F2 slope.

34

• 3. VOT has been the most frequently used index of
  subsystem coordination, because it is assumed
  that the acoustic interval between the burst and
  the onset of periodic energy corresponds to the
  physiological interval between release of the
  consonantal constriction and the onset of vocal
  fold vibration. Therefore, a large amount of
  data has been published on VOT in normal speech
  and several varieties of disordered speech.
• For Instance
• Longer VOTAtaxic Dysarthria
• More Variable VOTALS (Lou Gerhigs Disease)
• Shorter VOTSpastic Dysarthria

35

• 4. Fricatives are of interest bc their
  production requires precise articulatory control
  (i.e., to maintain the aperture that will
  generate frication noise). General
  characteristics of disordered speech production
  are longer segment durations and slower formant
  transitions at consonant-vowel boundaries.
• For Instance
• The spectrum shapes of /s/ and /S/ in initial
  position have been shown to be different between
  speech of people with and without ALS.

36
But this is a young fieldmore data are needed!

• No single acoustic measure, or even a small set
  of measures, is adequate of the purpose of
  describing all consonants. In fact, questions
  have been raised about the degree to which VOT in
  itself is a satisfactory index of the
  coordination of laryngeal and supralaryngeal
  events. In future research, it would be wise to
  supplement some other measures such as stop gap,
  voiceless interval, and aspiration to describe
  consonantal features.

37
Lab 4 Categorical Perception

• 8.30-9.45/12.30-1.45
• In pairs Speech Perception Speech Perception
  Module of Sensimetrics, SECTIONS 1-7
• (9.00 Hand Out Assignment)
• 9.45-10.15/1.45-2.15
• Group Discuss Research and Clinical Application
  of Acoustic Analysis
• a. Teaching Spectrograms to Kids example
• b. Compensatory Articulation in Kids example
• (c. VOT Dysarthria exs, time permitting, from
  Lab 3)
• 10.15-10.20/2.15-2.20
• QuestionsEspecially about the midterm?

38
How Well Can Children Recognize Speech Features
in Spectrograms? Comparisons by Age and Hearing
Status(Ertmer, 2004, JSLHR)

• Real-time spectrographic displays (SDs) have been
  used in speech training for more than 30 years
  with adults and children who have severe and
  profound hearing impairments. Despite positive
  outcomes from treatment studies, concerns remain
  that the complex and abstract nature of
  spectrograms may make these speech training aids
  unsuitable for use with children.
• This investigation examined how well children
  with normal hearing sensitivity and children with
  impaired hearing can recognize spectrographic
  cues for vowels and consonants, and the ages at
  which these visual cues are distinguished.

39
Abstract, continued

• 60 children (30HI, 30N) in 3 age groups (6-7,
  8-9, 10-11) were familiarized with the
  spectrographic characteristics of selected vowels
  and consonants. The children were then tested on
  their ability to select a match for a model
  spectrogram from among 3 choices.
• Overall scores indicated that spectrographic cues
  were recognized with greater-than-chance accuracy
  by all age groups. Formant contrasts were
  recognized with greater accuracy than consonant
  manner contrasts. Children with normal hearing
  sensitivity and those with hearing impairment
  performed equally well.

40
CHALLENGE QUESTION How would you teach
spectrograms to kids receiving speech treatment?

• Q How would you teach or explain spectrographic
  consonant and vowel features and cues to children
  in a way they could understand and remember?
• Think creatively!

41
What would you do? What did they do?
Nasals
Low freq. energy. Little shoe sounds Low
energy bar at the start of a word was equated
with the toe of a shoe sticking out from the
bottom of a pair of pants (increased energy
associated with formant onset)
Stops
Vertical line of energy across a wide range of
freq. Straight sounds vertical
burst was compared to the straight edge of a ruler
Fricatives
High freq. aperiodic energy. Fat Bushy sounds
The relatively broad and irregular shapr of
fricative energy was compared to a wide bush
42
Affricates
High freq. aperiodic energy. Skinny bushy
sounds The relatively narrow and irregular shape
of fricative energy was likened to a narrow bush
Glides
Changing location of formant energy. Sliders
Changes in formant location were likened to the
shape of a playground slide
Liquids
Changing location of formant energy. Crooked
sounds Changes in formant energy were described
as crooked, not perfectly horizontal
Vowels
Formant freq. location and relative distance.
Stripes in the middle of a word The locations
of F1 and F2 and the distance between than for a
variety of vowels were discussed. Analogies were
not given for specific vowel categories.
43
Example Stimuli (natural productions, 7yo boy)

• Target Choices
• kite might, night, kite
• dig fig, dig, zig
• beep cheap, jeep, beep
• thaw thaw, jaw, chaw
• yank rank, yank, lank
• hoot heat, hat, hoot
• bug bag, bog, bug
• heed heed, head, hid
• kook cook, kook, coke

44
Acoustic Analysis of Compensatory Articulation
in Children(Baum, et al, 1988, JASA)

• A study was undertaken to explore the effects of
  fixing the mandible with a bite block on the
  formant frequencies of the vowels i a u
  produced by two groups of children aged 4-5 and
  7-8 years. Vowels produced in both normal and
  bite-block conditions were submitted to acoustic
  analysis with windows placed over the first
  glottal pulse and at the vowel midpoint. For
  both groups of children, no differences were
  found in the frequencies of either the first of
  second formants between with normal and
  bite-block conditions.

45
CHALLENGE QUESTION

• Q This research is intended to inform us about
  the underlying theories of motor control
  acquisition. Since the compensated kids
  produced acoustically identical vowels as the
  normal kids, it suggests that even young
  children aim for an acoustic/perceptual goals,
  and not on particular articulatory postures, in
  producing a target.
• In other words, the physical placement of
  articulators can produce acoustically and
  perceptually equivalent speech sounds.
• What can this fact tell us about how clinicians
  should treat articulation disorders? Should
  clinicians just focus on articulatory instruction
  (ie shape your mouth like this)?

46

• A NO! It is especially difficult to give
  articulator placement instruction with vowels,
  since there is so much variability. Instead,
  clinicians can focus on using perceptual cues to
  shape acoustic production. And as we have just
  learned, training kids with spectrograms may be a
  helpful tool in shaping vocal behavior!

47
Lab 5

• 8.30-9.30/12.30-1.30
• Native Language Influences Experiment (Hindi)
• 930-1020/1.30-2.20
• 1) Discussions that we didnt get to cover in
  previous labs
• 2) Sensimetrics
• (Review Previous modules, or Vowel Perception
  module)
• 3) Questions

48
Hindi Experiment

• http//lrs.ed.uiuc.edu/Students/avatans/varna.html
  
• Hindi Consonants Section
• Articulatory Explanation of Placement
• Google Comparative Analysis of Hindi Retroflex
  Syllables ? view as html
• Intro Para. 2
• Sections 3.1-3.4 Conclusions Para. 1
• http//courses.washington.edu/SPHSC425/labs/lab5/l
  ab5_exp.htm
• Worksheet
• Google Perceptual Training on Hindi dental and
  retroflex consonants Pruitt

49
Lab 6 Audio-Visual Speech

• 8.30-9.30
• AV Speech Perception Experiment
• 930-1020
• Group Discuss Research and Clinical Application
  of AV Speech
• a. Aphasia example
• b. Oesophageal Speech example
• c. French Vowels example

50
McGurk Illusions

• Illusions occur with Incongruent pairings
• FUSIONS
• Visual Velar Aud Bilabial Alveolar
• /ga/ /ba/ /da/
• COMBINATIONS
• Visual Bilabal Aud Velar BilabialVelar
• /ba/ /ga/ /abga/

51
Auditory-visual speech perception in an adult
with aphasia Youse, et al., 2004, Brain Injury

• The evaluation of auditory-visual speech
  perception is not typically undertaken in the
  assessment of aphasia however, treatment
  approaches utilize bimodal presentations.
• Research demonstrates that auditory and visual
  information are integrated for speech perception.
  The strongest evidence of this cross-modal
  integration is the McGurk effect. This indirect
  measure of integration shows that presentation of
  conflicting tokens may change perception (e.g.
  auditory /bi/ visual /gi/ /di/).

52
Continued

• The purpose of this study was to investigate the
  ability of a person with mild aphasia to identify
  tokens presented in auditory-only, visual-only
  and auditory-visual conditions. It was
  hypothesized that performance would be best in
  the bimodal condition and that presence of the
  McGurk effect would demonstrate integration of
  speech information.
• Findings did not support the hypotheses. It is
  suspected that successful integration of AV
  speech information was limited by a perseverative
  response pattern. This case study suggests the
  use of bisensory speech information may be
  impaired in adults with aphasia.

53
Challenge Question

• Q Many therapists use lipreading training as
  part of treatment for clients with aphasia.
• What kinds of things might you need to find out
  first, to make sure this approach would be
  helpful?
• If you had a client like the one described in
  this paper, do you think lipreading training
  would help?

54

• A
• -Since this person did not seem to particularly
  benefit from AV speech, it isnt likely that
  lipreading would be more beneficial than other
  types of intervention.
• -Try meaningful words many people with aphasia
  have more trouble with non-words and syllables
  than with words this study only used abstract
  syllables.
• -Test comprehension through visual-only,
  auditory-only, then both, and see if performance
  is improved with added visual information. If
  not, lipreading might not be efficient.
• -Test incongruent items (like this exp) to see if
  there is successful cross-modal integration. If
  not, lipreading might not be efficient.

55
Auditory versus audio-visual intelligibility
measurements of alaryngeal speech A preliminary
report(Berry, et al., 1974, Perceptual and Motor
Skills)

• This investigation supports the notion that
  audio-visual presentations of esophageal speech
  to 32 university student judges yield a relative
  increase in rated intelligibility in contrast
  with esophageal speech presented auditorily.
  Implications suggest that to construct realistic
  therapeutic goals of an esophageal speaker more
  effectively, the audio-visual component should be
  included in the clinical assessment.

56
Challenge Question

• Q Do you think that there may be other types of
  conditions, besides alaryngeal speech, for which
  an AV assessment component could be appropriate?

A Aphasias (last example) children on the ASD
dysarthrias TBI Hearing-ImpairmentSeveral of
these areas have been studied, but not many as of
yet.
57
Effects of phonetic context on audio-visual
intelligibility of FrenchBenoart, et al., 1994.
JSHR

• Bimodal perception leads to better speech
  understanding than auditory perception alone. We
  evaluated the overall benefit of lip-reading on
  natural utterances of French produced by a single
  speaker.
• Eighteen French subjects with good audition and
  vision were administered a closed set
  identification test of VCVCV nonsense words
  consisting of three vowels i, a, y and six
  consonants b, v, z, 3, R, l. Stimuli were
  presented under both auditory and audio-visual
  conditions with white noise added at various
  signal-to-noise ratios.

58
Continued

• Identification scores were higher in the bimodal
  condition than in the auditory-alone condition,
  especially in situations where acoustic
  information was reduced. The auditory and
  audio-visual intelligibility of the three vowels
  i, a, y averaged over the six consonantal
  contexts was evaluated as well.

59
Continued

• Two different hierarchies of intelligibility were
  found. Auditorily, a was most intelligible,
  followed by i and then by y whereas visually
  y was most intelligible, followed by a and
  i.
• We also quantified the contextual effects of the
  three vowels on the auditory and audio-visual
  intelligibility of the consonants. Both the
  auditory and the audio-visual intelligibility of
  surrounding consonants was highest in the a
  context, followed by the i context and lastly
  the y context.

60
Challenge Question

• Q1 Why might y be the most visually robust?

Hint Watch someone pronounce /y/, /i/, /a/
A Lip rounding! In this vowel, lip gesture is
most highly constrained.
61
Q2 Why might one have the most difficulty in
identifying fricatives adjacent to y? (CV)
Hint Consider the visual robustness of this
vowel!
A With its salient lip rounding, this vowel most
strongly distorts the surrounding C context
anticipatory lip rounding can make initial
fricatives appear more similar than they would
appear in other vocalic contexts (e.g., i or
a).
62
Online Lab Survey

• Check your email
• Please consider participating

63
DONT FORGET!!!

• Lab 7 is the Head-Turn and ERP tour at ILABS.
  DONT COME HERE!
• Check email for directions.
• HAPPY THANKSGIVING!!