Embed Size (px)
Citation preview
Syllables and Stress
October 25, 2010
Practicalities• Some homeworks to return…
• Review session on Wednesday.
• Mid-term on Friday.
• Note: transcriptions have been posted to the course web page.
• Let’s check out the Boston English sample.
Review: Suprasegmentals• Last time, we learned that there were three kinds of
languages:
1. Tone languages (Chinese, Navajo, Igbo)
• lexically determined tone on every syllable or word
2. Accentual languages (Japanese, Swedish)
• the location of an accent is lexically marked.
3. Stress languages (English, Russian)
• it’s complicated
What is Stress?• Examples of stress in English:
(V) vs. (N)
(V) vs. (N)
• Phonetically, stress is hard to define
• I.e., it is hard to measure.
• It seems to depend on an interaction of three quantifiable variables:
• Pitch
• Duration
• Loudness
• And also: quality
Loudness
• How do we measure how loud a sound is?
• Recall: one parameter of a sinewave is its amplitude.
• Peak amplitude (for sound) is the highest sound pressure reached during a particular wave cycle.
peak-to-peak amplitude
Amplitude/Loudness Examples• The higher the peak amplitude of a sinusoidal sound, the louder the sound seems to be.
RMS amplitude• Peak-to-peak amplitude is sufficient for characterizing
the loudness of sinewaves, but speech sounds are more complex.
• Another method of measuring loudness:
root-mean-square (RMS) amplitude
• To calculate RMS amplitude:
1. Square the pressure value of the waveform at each point (sample) in the sound file
2. Average all the squared values
3. Take the square root of the average
RMS example
pressure 1 0.707 0 -0.707 -1 -0.707 0 0.707 1sample 1 2 3 4 5 6 7 8 9
• A small sampling of a “sinewave” has the following pressure values:
• It looks like this (in Excel):
RMS calculationspressure 1 0.707 0 -0.707 -1 -0.707 0 0.707 1sample 1 2 3 4 5 6 7 8 9
• To calculate RMS amplitude for this sound, first square the values of each sample:
• Then average all the squared values
(1 + .5 + 0 + .5 + 1 + .5 + 0 + .5 + 1) / 9 = 5/9 = .555
• Then take the square root of the average
• RMS amplitude = .745
square 1 0.5 0 0.5 1 0.5 0 0.5 1sample 1 2 3 4 5 6 7 8 9
Another example• What about the RMS amplitude of this sound wave?
•It looks like this (in Excel):
pressure 1 1 1 -1 -1 -1 1 1 1sample 1 2 3 4 5 6 7 8 9
More Complex Waveforms• The following waveforms all have the same peak-to-peak amplitude:
Intensity
• Two related concepts are acoustic power and intensity.
• Power is just the square of amplitude.
• P = A2
• The intensity of a sound is its power relative to the power of some reference sound.
• Intensity is usually measured in decibels (dB).
• Decibels is a measure of intensity with reference to the quietest sound human ears can hear.
Some Numbers• The intensity of a sound x can be measured in bels, where a bel is defined as:
= log10 (x2 / r2)
• r2 is the power of the reference sound
• x2 is the power of sound x.
• A decibel is a tenth of a bel.
• Some typical decibel values:
30 dB Quiet library, soft whispers
40 dB Living room, refrigerator
50 dB Light traffic, quiet office
60 dB Normal conversation
Numbers, continued• Some typical decibel values:
70 dB Vacuum cleaner, hair dryer
80 dB City traffic, garbage disposal
90 dB Subway, motorcycle, lawn mower
100 dB Chain saw, pneumatic drill
120 dB Rock concert in front of speakers, thunderclap
130 dB Pain threshold
140 dB Gunshot blast, jet plane
180 dB Rocket launching
Intensity Interactions• Perceived loudness depends on frequency, as well as amplitude.
• Mid-range frequencies sound louder than low or extremely high frequencies.
• 100 Hz
• 250 Hz
• 440 Hz
• 1000 Hz
• 4000 Hz
• 10000 Hz
An Interesting Fact• Some vowels are louder than others
• dB of different vowels relative to (Fonagy, 1966):
: 0.0
[e] : -3.6
[o] : -7.2
[i] : -9.7
[u] : -12.3
• Why?
Another Interesting Fact• Some vowels are inherently longer than others.
• Data from Swedish (Elert, 1964):
long short
high [i y u] 140 msec 95
mid 155 103
low 164 111
• Why?
Sonority• Loudness is also a highly context-dependent measure.
• Can vary wildly within speaker, from speaker to speaker, from room to room, and across speaking contexts.
• However, all things being equal, some speech sounds are louder than others.
• Course in Phonetics:
“The sonority of a sound is its loudness relative to that of other sounds with the same length, stress and pitch.”
From Ladefoged
A Sonority Scale
low vowels
high vowels
glides
liquids
nasals
fricatives
stops
high sonority
low sonority
Sonority and Syllables
• An old idea (e.g., Pike, 1943): syllables are organized around peaks in sonority.
• This is the Sonority Sequencing Principle (SSP).
• Example: [bæd] is a well-formed syllable in English.
[æ]
[b] [d]
high sonority
low sonority
Sonority and Syllables
• An old idea (e.g., Pike, 1943): syllables are organized around peaks in sonority.
• This is the Sonority Sequencing Principle (SSP).
• Example: [blænd] works well, too.
[æ]
[l] [n]
[b] [d]
high sonority
low sonority
Technical Terms
[æ]
[l] [n]
[b] [d]
high sonority
low sonority
sonority peak
Technical Terms• The sonority peak forms the nucleus of the syllable.
[æ]
[l] [n]
[b] [d]
high sonority
low sonority
nucleus
Technical Terms• The sonority peak forms the nucleus of the syllable.
• The sounds that precede the nucleus form the syllable onset.
[æ]
[l] [n]
[b] [d]
high sonority
low sonority
onset
Technical Terms• The sonority peak forms the nucleus of the syllable.
• The sounds that precede the nucleus form the syllable onset.
• The sounds that follow the nucleus form the syllable coda.
[æ]
[l] [n]
[b] [d]
high sonority
low sonority
coda
Technical Terms• The sonority peak forms the nucleus of the syllable.
• The sounds that precede the nucleus form the syllable onset.
• The sounds that follow the nucleus form the syllable coda.
• Together, the nucleus and coda form the syllable rhyme.
[æ]
[l] [n]
[b] [d]
high sonority
low sonority
rhyme
Some basic principles• Onsets must rise in sonority towards the syllable peak.
Examples:
stop - {liquid/glide} ‘play’ ‘quick’
fricative - {liquid/glide} ‘fling’ ‘thwack’
[s] - {liquid/nasal/glide} ‘slide’ ‘snow’ ‘sweet’
• What onset clusters should be ruled out?
• Can you think of any English examples where this principle might not work?
Some basic principles• Codas must drop in sonority away from the syllable peak.
Examples:
nasal - {fricative/stop} ‘tenth’ ‘hand’
liquid - {fricative/nasal/stop} ‘help’ ‘helm’ ‘heart’
fricative - stop ‘test’
• What coda clusters should be ruled out?
• Can you think of any English examples where this principle might not work?
Other Problems• The Sonority Sequencing Principle doesn’t always work.
• How can we define a “syllable”?
• An alternative idea: each syllable is a “chest pulse” (Stetson, 1951)
• It turns out this doesn’t work, either.
• Chest muscles don’t necessarily contract for each syllable (Ladefoged, 1967)
• Any better ideas?
Interesting Patterns• Check out the following words:
Atlantic atrocious
America arcade
astronomy arthritic
• When is the first vowel a ?
• Is there a difference between the /t/ in ‘atrocious’ and the /t/ in ‘Atlantic’?
• Why?
Syllables “defined”• “Syllables are necessary units in the organization and
production of utterances.” (Ladefoged, 1982)
• The construct of a “syllable” can account for a variety of interesting phonological patterns:
1. Vowel reduction in unstressed syllables in English.
2. Fricatives and stops devoice at the end of syllables in German, Russian (and many other languages)
3. Place contrasts disappear in coda position in Japanese (and many other languages)
4. Voiceless stops are aspirated at the onset of stressed syllables in many Germanic languages.
Back to Stress• Stress is a suprasegmental property that applies to whole syllables.
• Stressed syllables are higher in pitch (usually)
• Stressed syllables are longer (usually)
• Stressed syllables are louder (usually)
• Stressed syllables reflect more phonetic effort.
• The combination of these factors give stressed syllables more prominence than unstressed syllables.
Stress: Pitch
(N)
Complicating factor: pitch tends to drift downwards at the end of utterances
(V)
Stress: Intensity
(N)
(V)
Perception of stress is highly correlated with the area under the intensity curve
“Phonetic Effort”• Voiceless stops are more aspirated at the onset of
stressed syllables in English
• Vowels are often reduced to in unstressed syllables in English.
• There is less coarticulation across syllable boundaries in stressed syllables.
• X-Ray microbeam study (deJong et al., 1993); two utterances:
1. I said put the TOAST on the table, not the napkins!
2. I said PUT the toast on the table, don’t throw it!
X-ray microbeam data
Varying Levels of Stress/Prominence
• English has both primary and secondary stress.
• Example: “exploitation”
vowel X X X X
full vowel X X X
stress X X
tonic accent X
Open Questions• Stress appears to be a property of whole syllables.
• What’s a syllable?
• It reflects overall phonetic effort.
• What’s that?
• Stressed syllables are louder than unstressed syllables.
• What’s loudness?
• Loudness is the easiest of these concepts to define…
