Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

LING 333 > Quiz 3 Transcription + > Flashcards

Quiz 3 Transcription + Flashcards

1
Q

What factors determine how much of the recorded audio gets transcribed?

A
  • what serves the purpose (interviewer’s speech might not be of interest, carrier phrase may be excluded ex. “Say __ again”
  • how much time at hand, how much data to transcribe
  • how much money to hire transcribers
  • quality of audio
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is the issue with using MS-Word to transcribe?

A
  • does not align the transcription with the audio
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is ELAN?

A
  • a powerful, popular software tool for transcribing data
  • especially useful when you want to transcribe video data
  • allows multiple tiers
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

How long does transcribing take, on average?

A

1 hour for 6-8 minutes of speech

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

Which software is ideal for transcribing video data?

A

ELAN

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

Describe Praat

A
  • one of the easiest ways to transcribe audio data
  • unlike word processors, transcriptions are time-aligned with audio
  • allows use of multiple tiers to include different levels of annotations
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What are the symbol systems used in transcription?

A
  • IPA
  • SAMPA
  • ARPAbet
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What does SAMPA stand for?

A

Speech Assessment Methods Phonetic Alphabet

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

Why was SAMPA developed?

A
  • to allow for transcription in electronic text formats
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What are the key features of SAMPA?

A
  • uses standard ASCII characters to represent phonetic symbols (extension of IPA, so not just American English)
  • designed to simplify IPA symbols for easy use in computational systems
  • directly replaces IPA symbols with ASCII keys
  • widely used in computational linguistics and speech synthesis applications
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Why was SAMPA created?

A
  • computers and earlier software systems had limited ability to handle IPA symbols
  • SAMPA provided a workaround using basic text, ensuring compatibility across different systems
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What does ARPAbet stand for?

A

Advanced Research Projects Agency for speech processing

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What is ARPAbet?

A
  • a phonetic transcription system developed by the Advanced Research Projects Agency for speech processing
  • like SAMPA, uses ASCII characters to represent phonemes, but is specifically designed for use in automatic speech recognition and speech synthesis
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What are the key features of ARPAbet?

A
  • uses a distinct set of ASCII symbols for each phoneme, based on American English
  • Each phoneme represented by capital letters and numbers (ex. /k/ represented as K)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Why is ARPAbet used?

A
  • simplifies the representation of English phonemes for use in speech technology
  • easy to integrate into speech synthesis and recognition software without needing special characters
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What are formants?

A
  • resonance frequencies of our vocal tract
17
Q

Describe the frequencies generated by the vocal folds

A
  • vocal folds generate a lot of frequencies
  • some get amplified, others dampened by our vocal tract (a tube)
  • amplified frequencies are formant frequencies
  • one tube can have multiple formant frequencies
18
Q

Which formant frequencies of vowels are used for phonetic analysis?

A
  • mostly just the first 2 or (at best) 3
19
Q

What determines whether a frequency is dampened or amplified?

A

When the frequency of the sound waves produced by the vocal folds match the resonant frequencies of the vocal tract, you get amplification. And when they don’t match, you get dampening.

20
Q

When using forced alignment for vowel analysis, is it preferable to use stressed or unstessed vowels?

A

Stressed is better as unstressed are less consistent and alignment doesn’t work as well

21
Q

To see isolated formants zoomed in, where do you go in Praat?

A

Spectrogram>view spectral slides
Peaks are formants (places where frequency is boosted)—frequency is on X axis
1st peak= F1 (click on peak to see frequency)
F0- pitch (fundamental frequency)

22
Q

At which formant is rhoticity usually measured?

A

F3

23
Q

What can F1 tell us about a vowel? F2?

A

F1
- corresponds to vowel’s openness (“height”)
- lower values indicating a more closed (“high”) vowel ex. /i/
- higher values indicating a more open (“low”) vowel ex. /a/
F2
- corresponds to frontness/backness of the tongue
- higher values indicating a more front vowel ex. /i/
- lower F2 values = more back ex. /u/

24
Q

What is the name for a spectrum that is more smoothed out and less detailed?

A

Wideband/broadband

25
Q

What is the advantage of a broadband spectrum as opposed to a narrowband?

A

Easier to detect formants

26
Q

What is the advantage of using a narrowband spectrum?

A

It shows every frequency of the vocal folds (the “source”)

27
Q

How is the fundamental frequency (F0) determined

A

From very first frequency
- corresponds to H1 and = pitch

28
Q

What are all of the individual frequencies on a narrowband spectrum called?

A

Harmonics (H1, H2, H3, etc)

29
Q

What kind of spectrum do you need to see the F0?

A

Narrowband

30
Q

What is another term for harmonics?

A

Component frequencies

31
Q

How to create narrowband vs broadband spectrum?

A

Spectrogram> settings> window length (.005 default is broadband, .030 narrowband)

5ms vs 30ms (broad vs narrow)

32
Q

How to read spectrograms?

A

On spectrogram, each vertical line corresponds to spectrum
*spectrum will not show any time information
*darker bands= higher amplitude

33
Q

What can you do to confirm where formants are on a spectrogram in Praat?

A

Formants > show formants
(Not perfect at identifying, though)

34
Q

If Praat can’t differentiate between formants, what can you do to better see?

A

Spectrogram settings: can change Show frequencies up to ________by reducing “view range” Hz

Formant settings>number of formants (can increase from default)

35
Q

What is the F1/F2 combo for /i/?

A

High, front vowel
Low F1, High F2

36
Q

What is the F1/F2 combo for /u/?

A

High, back vowel
- low F1, low F2

37
Q

What is the F1/F2 combo for /o/?

A

Slightly low F1, low F2

38
Q

What are the 3 (acoustic) components of vowels that can be measured?

A
  • duration
  • formants
  • phonation
39
Q

What is on each axis of a frequency spectrum?

A

Y- intensity/amplitude, measured in dB/Hz
X- frequency, measured in Hz

LING 333 (11 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions