L5 - Voice and Speech Processing Flashcards

1
Q

Mcleer et al 2014: (voice perception study)

A
  • Ps rated social traits on ‘hello’ spoken by 64 people
  • listen to hello and rate on 10 different traits
  • if voices convey no info, ratings would be random
  • but CONBACH ALPHA rating - suggests that people CONSISTENTLY AGREE on whether a voice sounds attractive or confident
  • we seem to associate voices with social traits on a collective level
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

many social traits are related and can be mapped to two principle components:

A

in both genders:
1. the 1st principle component is highly correlated with TRUSTWORTHINESS

  1. the 2nd principle component is highly correlated with DOMINANCE.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

how are voices transmitted?

A

transmitted via vibrations of the air

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

three main parameters of the voice

A
  • pitch
  • intensity
  • tempo
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

different cultures?

A

different cultures may struggle to determine emotion in voice when from very different language.

e.g. Indonesian struggle with German

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

area of the brain processing voices?

A

temporal voice area (TVA)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Belin et al 2000

A

found that when P listens to voices versus non-vocal sounds there was greater bilateral activity in the temporal cortices

  • evidence that we possess specific brain regions for voice processing
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

neural adaptation

A

a gradual decrease over time in the responsiveness of the sensory system to a constant stimulus

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

speaker identity and vocal emotion?

A

processed in different regions of the TVA

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

speech processing - top down?

A

speech is a continuous signal that doesn’t have punctuation or space - how does the brain extract this?

the brain determines what we see/hear and how we perceive follows one of two processing methods.

  • our brain extracts LINGUISTIC UNITS of speech such as syllables and changes in tone
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

cocktail party problem

A
  • once the brain knows what to look for in disordered speech - it cannot unhear it (top-down)
  • this is why you can hear someone speaking in a club - visually attended
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

what side of the brain is more responsible for speech?

A

left

speech is left lateralised

  • evidence of this comes from neuropschyology - BROCAS and WERNICKES
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Wernickes

A
  • patient cant understand speech
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Brocas

A
  • patient cannot produce speech
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

left temporal regions

A

more specialised in processing time information

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

right temporal regions

A

more specialised in processing spectral information

17
Q

acoustic information (music)

A
  • musical
  • pitch more important
  • therefore SPECTRAL info MORE important than TEMPORAL info
18
Q

speech - temporal or spectral?

A

more important to have TEMPORAL then SPECTRAL information

  • as matters more regarding timing rather then gender
19
Q

the acoustic uncertainty principle

A

you need time to get the frequency correct.

the principle is that there is a trade-off between time and frequency:

  • as that - the upper limit of resolution, increasing time resolution can only be achieved at the expense of frequency (spectral) resolution and vice versa
20
Q

the McGurk Effect

A
  • voice saying ‘ba ba’

is heard as: ‘ba ba’ ‘da da’ or ‘va va’
- depending what you are perceiving at the time by looking at mouth movement

21
Q

motor theory of speech perception

A
  • the auditory signal is matched on to motor representations for producing ones own speech rather than matching to an acoustic template
22
Q

neural oscillations for speech segmentation

A
  • may be modulated by the motor system
  • facilitates the segmentation of speech signal
  • optimal oscillations = 4.5Hz which reflects average syllable rate across human languages
23
Q

why does the McGurk effect arise ?

A
  • one theory could be there is a multi-sensory perception of speech - with involvement of the left superior temporal region
  • another theory is that of motor system activation - involving the inferior frontal cortex/premotor regions
24
Q

alpha oscillations?

A

linked to suppression of irrelevant information

25
Q

speech is processed in a network

A

dorsal ‘how’

ventral ‘what’

26
Q

auditory-motor connectivity

A
  • McGurk Illusion

- Neural oscillations

27
Q

superior temporal sulcus is specialised for…

A

processing voice over nonvocal words