accoustic evidence for uninterruption of speech
dung...@yahoo.com
continously. In speech the prior sentence would be:
Thoughleavingspacebetweenwordsinwriting, speechisutteredcoutinuously.
I heard there are machines to record the flow of speech graphically, so
people can 'see' it. In fact I've seen such graphs, but wonder what the
name of the machine or accoustic device is?
Thanks.
Alexander Magidow
I did a lot of work with looking at sounds(ie acoustic phonetics), and
we just used computer with microphones- the sounds are just recorded as
.wav(or a proprietary format if we recorded in our usual program,
PCQuirer(sp?)), then a program to look at the sound in various
ways(waveform, spectrogram, etc), of which there are several available
for free.
Here's an example of a spectrogram:
http://en.wikipedia.org/wiki/Image:Praat-spectrogram-tatata.png
dung...@yahoo.com
Not only that type. There are 4 graphs in following page. My
questions are in it, too. Please visit:
http://www.pinyinology.com/gr/zhangfile.html
Thanks.
dung...@yahoo.com
two. So 'acoustic'.
Tommi Nieminen
> There are 4 graphs in following page. My
> questions are in it, too. Please visit:
>
> http://www.pinyinology.com/gr/zhangfile.html
1. Wave format (time/amplitude).
2. Without any further information such as the units of the coordinates,
could be almost anything! For example, a fundamental frequency or
intensity graph. (But since the graph is higher during the consonants
than during the vowel...)
3. Looks very much like the same as (1)!
4. Spectrogram (time/frequency/intensity). Also known as "Sonagram"
after the firm that used to produce specific machines creating these
graphs (before the computer era).
Each and every one of those graphs could produced with exactly the same
software, such as Praat <http://www.praat.org/>.
You really should first get yourself a copy of a decent introductory
book on phonetics, such as:
Borden, Gloria J. & Harris, Katherine S. & Raphael, Lawrence J.: Speech
Science Primer. Physiology, Acoustics, and Perception of Speech. Fourth
edition. Philadelphia: Lippincott, Williams and Wilkins 2003.
--
.... Tommi Nieminen .... http://www.saunalahti.fi/~tommni/ ....
Csak füvön élt a kis zebra,
de most rákapott a zabra;
végül is elvitték Szobra,
ott oktatják szebbre-jobbra.
-Devecseri Gábor-
.... tommi dot nieminen at campus dot jyu dot fi ....
Alexander Magidow
> dung...@yahoo.com kirjoitti:
>
>> There are 4 graphs in following page. My
>> questions are in it, too. Please visit:
>>
>> http://www.pinyinology.com/gr/zhangfile.html
>
>
> 1. Wave format (time/amplitude).
>
> 2. Without any further information such as the units of the coordinates,
> could be almost anything! For example, a fundamental frequency or
> intensity graph. (But since the graph is higher during the consonants
> than during the vowel...)
>
> 3. Looks very much like the same as (1)!
>
> 4. Spectrogram (time/frequency/intensity). Also known as "Sonagram"
> after the firm that used to produce specific machines creating these
> graphs (before the computer era).
>
> Each and every one of those graphs could produced with exactly the same
> software, such as Praat <http://www.praat.org/>.
>
> You really should first get yourself a copy of a decent introductory
> book on phonetics, such as:
>
> Borden, Gloria J. & Harris, Katherine S. & Raphael, Lawrence J.: Speech
> Science Primer. Physiology, Acoustics, and Perception of Speech. Fourth
> edition. Philadelphia: Lippincott, Williams and Wilkins 2003.
>
probably already has such a book...
Dylan Sung
"Alexander Magidow" <amag...@wisc.edu> wrote in message
news:_9sEe.3807$Nh3....@tornado.rdc-kc.rr.com...
No it's our good friend Juli Zhang who promotes pinyinisation of Chinese.
The second graph looks like sometime to do with pitch contours, perhaps
Peter will know. The third one is indeed the first one to a different scale.
Dyl.
Tommi Nieminen
> The second graph looks like sometime to do with pitch contours,
Could be, though it's strange in its way. Not real-time, I guess--maybe
filtered somehow?
--
... Tommi Nieminen ... http://www.saunalahti.fi/~tommni/ ...
O Gracious God! how far have we
Profaned thy heavenly gift of poesy!
-John Dryden-
... tommi dot nieminen at campus dot jyvaskyla dot fi ...
Alexander Magidow
> Dylan Sung kirjoitti:
>
>> The second graph looks like sometime to do with pitch contours,
>
>
> Could be, though it's strange in its way. Not real-time, I guess--maybe
> filtered somehow?
>
does appear to be a plot of frequency versus time.
Tommi Nieminen
> It looks a lot like fundamental frequency graphs that I've seen,
"Looks a lot like" can hardly be considered evidence! F0 and A0 graphs,
for example, look very similar. The "strange thing" I mentioned is that
the graph is *higher* during the consonants; not totally impossible, but
raises doubts. Further, unfiltered F0 graphs tend to change more rapidly
and have gaps in them.
--
.... Tommi Nieminen .... http://www.saunalahti.fi/~tommni/ ....
Csak füvön élt a kis zebra,
de most rákapott a zabra;
végül is elvitték Szobra,
ott oktatják szebbre-jobbra.
-Devecseri Gábor-
.... tommi dot nieminen at campus dot jyu dot fi ....
Bart Mathias
> Alexander Magidow kirjoitti:
>
>> It looks a lot like fundamental frequency graphs that I've seen,
>
>
> "Looks a lot like" can hardly be considered evidence! F0 and A0 graphs,
> for example, look very similar. The "strange thing" I mentioned is that
> the graph is *higher* during the consonants; not totally impossible, but
> raises doubts. Further, unfiltered F0 graphs tend to change more rapidly
> and have gaps in them.
If it's Mandarin second or third tone, I think the "ng" part would
naturally be higher, and, if third, at least, probably the "zh" part,
too. I would expect things to be set to give more "depth" though.
Bart Mathias
Dylan Sung
"Bart Mathias" <mat...@hawaii.edu> wrote in message
news:20yEe.5043$dM3.1358@trnddc04...
tone 1 /55/
tone 2 /35/
tone 3 /213/ or /214/
tone 4 /51/
My guess is that the 'zhang' is of Juli Zhang's surname, the character
formed of the radicals for bow and long. It would be in tone 1, though.
Dyl.
dung...@yahoo.com
I requested the syllable be in first tone. The production was a sound
in female voice. It sounded a firt-tone syllable to my ear. But the
file containing the sound has been lost.
dung...@yahoo.com
mandarin is different from that of non-tone languages. Probably there
is a brief silence between the tonal syllables, so that the speaker can
switch from one tone pattern to another.
Please note, just 'probably'. If anyone knows research in this field,
information is greatly appreciated.
Lee Sau Dan
dungping5> Becasue of the tones, I expect the relationship between
dungping5> syllables in mandarin is different from that of
dungping5> non-tone languages.
How? And why?
dungping5> Probably there is a brief silence between the tonal
dungping5> syllables, so that the speaker can switch from one tone
dungping5> pattern to another.
Why would that be NECESSARY? Don't you need pauses, then, between
sound segments in non-tonal languages to allow the speaker to switch
from one sound segment to another? e.g. don't you need a pause
between "s" and "l" in "slim" to allow the speaker to switch from the
voiceless "s" to the voiced "l"? And you'd need a pause to switch
from the "s" in "sky" to the "k" so as to switch the articulation
point?
--
Lee Sau Dan 李守敦 ~{@nJX6X~}
E-mail: dan...@informatik.uni-freiburg.de
Home page: http://www.informatik.uni-freiburg.de/~danlee
dung...@yahoo.com
You talk about elements in single English syllables, 'slim' and 'sky'.
Furthermore, I don't say there is pause in non-tonal languages.
I talk about investigating whether there is pause between Chinese
syllables with modern technology. The hypothesis is the speaker has to
change tone. If there is, then answering the question 'why'.
You are totally off the topic. Meaningful discussion is not possible
because of you.
Lee Sau Dan
dungping5> I say there may be pause between neighboring full
dungping5> Chinese syllables. You talk about elements in single
dungping5> English syllables, 'slim' and 'sky'. Furthermore, I
dungping5> don't say there is pause in non-tonal languages.
Because I don't think a tone language is that different from non-tone
languages in that aspect.
dungping5> I talk about investigating whether there is pause
dungping5> between Chinese syllables with modern technology. The
dungping5> hypothesis is the speaker has to change tone. If there
dungping5> is, then answering the question 'why'.
There is nothing that is "change tone". The "tone" is just an
abstract concept at the phonological level. In phonetics, it's a
contour of pitch levels. Better described with a 2-D pitch vs. time
graph. In Mandarin, for instance, a "tone" is realized as a contour
of CHANGING pitch levels (except for tone #1, whose pitch level stays
_conceptually_ constant). e.g. for tone #2, the pitch rises middle
level to high level. So, I don't think there is any special "tone
change" to occur between syllable boundaries. The pitch level changes
along the whole utterance without interruption, even for non-tonal
languages. (Unless you can cite a language whose speakers speak in a
monotone.)
OTOH, Vietnamese tone "~" does have an interruption (glottal stop) *in
the middle* of that syllable. That interruption is a part of the tone
contour.
Nigel Greenwood
> dungping5> I talk about investigating whether there is pause
> dungping5> between Chinese syllables with modern technology. The
> dungping5> hypothesis is the speaker has to change tone. If there
> dungping5> is, then answering the question 'why'.
>
> There is nothing that is "change tone". The "tone" is just an
> abstract concept at the phonological level. In phonetics, it's a
> contour of pitch levels. Better described with a 2-D pitch vs. time
> graph. In Mandarin, for instance, a "tone" is realized as a contour
> of CHANGING pitch levels (except for tone #1, whose pitch level stays
> _conceptually_ constant). e.g. for tone #2, the pitch rises middle
> level to high level. So, I don't think there is any special "tone
> change" to occur between syllable boundaries. The pitch level changes
> along the whole utterance without interruption, even for non-tonal
> languages.
One aspect so far missing from this discussion is Tone Sandhi. In
Mandarin the best-known case is 3rd tone + 3rd tone => 2T + 3T: eg
shui3guo3 (fruit) => shui2guo3. YR Chao in his Mandarin Primer also
identified cases of sandhi when two 2T come together in three-syllable
phrases (the second changing to 1T): eg dong1nan2-feng1 (SE wind) =>
dong1nan1feng1.
Presumably this phenomenon arises because of the difficulty of
preserving the citation-form tones in continuous speech.
Nigel
--
ScriptMaster language resources (Chinese/Modern & Classical
Greek/IPA/Persian/Russian/Turkish):
http://www.elgin.free-online.co.uk
Nigel Greenwood
I wrote:
> YR Chao in his Mandarin Primer also
> identified cases of sandhi when two 2T come together in three-syllable
> phrases (the second changing to 1T): eg dong1nan2-feng1 (SE wind) =>
> dong1nan1feng1.
Sorry, it isn't just two 2T coming together -- it also applies to 1T/2T
+ 2T + any tone (as in the example I gave). In fact, because of the 3T
+ 3T rule ( => 2T + 3T) it also applies to phrases such as wo3 ye3 you3
(I also have) => wo2 ye1 you3.
dung...@yahoo.com
Lee Sau Dan wrote:
> >>>>> "dungping5" == dungping5 <dung...@yahoo.com> writes:
>
> dungping5> I say there may be pause between neighboring full
> dungping5> Chinese syllables. You talk about elements in single
> dungping5> English syllables, 'slim' and 'sky'. Furthermore, I
> dungping5> don't say there is pause in non-tonal languages.
>
> Because I don't think a tone language is that different from non-tone
> languages in that aspect.
>
>
> dungping5> I talk about investigating whether there is pause
> dungping5> between Chinese syllables with modern technology. The
> dungping5> hypothesis is the speaker has to change tone. If there
> dungping5> is, then answering the question 'why'.
>
> There is nothing that is "change tone". The "tone" is just an
> abstract concept at the phonological level.
Not only the tone, but also the whole sound, is abstract. Following a
correspondent's suggestion, I've checked out the Speech Science Primer,
but only the 3d ed. The 4th ed. has been ordered, but shipping takes
time. Somewhere it reads:
'The first thing to understand about sound is that it has no substance;
it is not a thing. It has no mass or weight, but it is rather a set of
movements or a disturbance. A sound wave can exist as a disturbance in
a gas such as air . . .'
Different tones in mandarin must make different disturbance in air,
similar to those by different phonomes, to make sense for both the
speaker and listener. I don't mean tones are same as phonemes. But the
difference is not abstract or non-abstract.
The relation between syllables in mandarin, a tone language, should be
different from that of non-tone languages. An evidence is the
impossibility of compounding in mandarin. In English, when 'black' and
'board' become 'blackboard', the 'board' becomes an unstressed
syllable. Such process does not occur in mandarin, because each
syllable has a tone. So the'blackboard' is pronounced exactly same as
'black board'. Also, the pronunciations of disyllabic words in
mandarin, if there are any, are same as they consist of monosyllabic
words.
I have to read the following part of your message more times before
responding.
Peter T. Daniels
Even back when I was playing with the spectrograph, you could flip a
switch for two different graphs. Presumably these electronic ones have
dozens of different settings.
(In the olden days, a brief utterance was painted electrically onto a
vertical cylinder, I suppose the way a Xerox machine drum works, with
coated paper wrapped around it, and an arm with a spark plug rose as the
cylinder rotated, burning the varying charge into the paper. The x-axis
is time, the y-axis frequency, and the darkness of the trace is
amplitude. This is the setting that shows the formants clearly; the
other setting plots a summed waveform that's less useful for studying
speech specifically. You can see the results in a 1948 book called, like
several others, *Visible Speech*, of which Martin Joos was an author; it
was a government document [classified during the war] reprinted in
hardcover by Dover. Or, of course, any phonetics textbook, but it'll
only have a few examples.)
--
Peter T. Daniels gram...@att.net
Tommi Nieminen
> Even back when I was playing with the spectrograph, you could flip a
> switch for two different graphs.
Probably broadband and narrowband spectrographs. Broadband spectrographs
are what are commonly called "spectrographs" nowadays: they are suitable
for segmental analysis. Narrowband spectrographs provided a means for F0
analysis, but there's many much better ways of doing that nowadays.
> Presumably these electronic ones have
> dozens of different settings.
You bet!
Lee Sau Dan
Nigel> I wrote:
>> YR Chao in his Mandarin Primer also identified cases of sandhi
>> when two 2T come together in three-syllable phrases (the second
>> changing to 1T): eg dong1nan2-feng1 (SE wind) =>
>> dong1nan1feng1.
Nigel> Sorry, it isn't just two 2T coming together -- it also
Nigel> applies to 1T/2T + 2T + any tone (as in the example I
Nigel> gave). In fact, because of the 3T + 3T rule ( => 2T + 3T)
Nigel> it also applies to phrases such as wo3 ye3 you3 (I also
Nigel> have) => wo2 ye1 you3.
Which Mandarin dialects are you talking about? I've never observed
these in Putonghua. Only the 3T + 3T ==> 2T + 3T is observed in PTH.
Of course, a natural question to ask about this rule is: what happens
when more than 2 3T occurs in a row? In some cases, the result is 3T
+ 2T + 3T (compare how people say "year 1997" and "year 1999".) In
others, it is 2T + 2T + 3T.
Lee Sau Dan
dungping5> The relation between syllables in mandarin, a tone
dungping5> language, should be different from that of non-tone
dungping5> languages. An evidence is the impossibility of
dungping5> compounding in mandarin. In English, when 'black' and
dungping5> 'board' become 'blackboard', the 'board' becomes an
dungping5> unstressed syllable. Such process does not occur in
dungping5> mandarin,
It does!
dungping5> because each syllable has a tone.
Are you aware of the "neutral tone" <qing1sheng1>?
How would you pronounce <yi1fu>, <jiao3zi>, <jie3jie>, etc. without
"unstressing" the second syllable? Not also that the neutral tone
affects (disables) the tone-sandhi rule: 3T+3T => 2T+3T. So, <jiao3>
+ <zi3> is not pronounced as the sandhied <jiao2zi3>. Rather, the
sandhi-ing is inhibited because the <zi3> suffix isn't really, <zi3>,
but <zi>. So, the result of this compound is <jiao3zi>. No
sandhi-ing. <zi> is basically "unstressed" in this case.
dungping5> So the'blackboard' is pronounced exactly same as 'black
dungping5> board'.
"is" or "is not"?
dungping5> Also, the pronunciations of disyllabic words in
dungping5> mandarin, if there are any, are same as they consist of
dungping5> monosyllabic words.
Explain to me <yi1fu>, <jiao3zi>. Would the words <fu4> and <zi3>
pronounced with the neutral tone when they're used STANDALONE?
Nigel Greenwood
Lee Sau Dan wrote:
>
> Which Mandarin dialects are you talking about? I've never observed
> these in Putonghua. Only the 3T + 3T ==> 2T + 3T is observed in PTH.
I assume that Chao, whom I was quoting, was referring to informal
Peking/Peiping/Beijing/Beeijing speech. He's at pains to add that this
rule "applies only to speech at conversational speed. In very
deliberate speech or slow reading, the 2nd Tone is unchanged."
In his magisterial 1968 A Grammar of Spoken Chinese, Chao refers to the
modification of 2nd Tone as "a tone sandhi of minor importance ...
speech at conversational speed". He gives a dozen or so examples,
including:
Sheir neng fei? [shei2 neng2 fei1] => 211 ('who can fly')
Hair mei wan [hai2 mei2 wan2] => 212 ('not yet finished')
Haojii-joong [hao3ji3 zhong3] => 213 ('quite a few kinds')
I doubt whether he was imagining this phenomenon.
dung...@yahoo.com
as variant of zi3 for that morpheme. Both jiao3zi and jiao3zi3 are
correct. Jiao3zi is from the Beijing dialect. Its original sound is
jiao3zi3 which is very popular in other mandarin dialects. When
non-mandarin speakers say this word in mandarin, the sound is jiao3zi3,
too.
As same morpheme, the zi must be in full 3d tone in following words:
zhang3 zi3 (eldest son), ci4 zi3 (second son), qi2 zi3 (chessman), bi4
ma2 zi3 (castor bean), kui2 hua1 zi3 (sunflower seed), fen1 zi3
(numerator of a fraction).
I don't speak Vietnamese, but verture to say that there is no
compounding in Vietnamese, either, and the reason is the tones, too.
Lee Sau Dan
Nigel> I assume that Chao, whom I was quoting, was referring to
Nigel> informal Peking/Peiping/Beijing/Beeijing speech.
So, it's not Putonghua. It's Pekingese.
Nigel> In his magisterial 1968 A Grammar of Spoken Chinese, Chao
Nigel> refers to the modification of 2nd Tone as "a tone sandhi of
Nigel> minor importance ... speech at conversational speed". He
Nigel> gives a dozen or so examples, including:
Nigel> Sheir neng fei? [shei2 neng2 fei1] => 211 ('who can fly')
Nigel> Hair mei wan [hai2 mei2 wan2] => 212 ('not yet finished')
Nigel> Haojii-joong [hao3ji3 zhong3] => 213 ('quite a few kinds')
Never heard these non-standard sandhi's in Putonghua.
Nigel> I doubt whether he was imagining this phenomenon.
Maybe not. Pekingese is different from Putonghua.
Bart Mathias
> Bart Mathias wrote:
> [..]
>>
>>If it's Mandarin second or third tone, I think the "ng" part would
>>naturally be higher, and, if third, at least, probably the "zh" part,
>>too. I would expect things to be set to give more "depth" though.
>
> I requested the syllable be in first tone. The production was a sound
> in female voice. It sounded a firt-tone syllable to my ear. But the
> file containing the sound has been lost.
First tone would probably explain the lack of depth I remarked. Even a
level tone is not likely to be like a sung monotone in normal human speech.
Bart Mathias
dung...@yahoo.com
syllables are uttered one after the other, the first syllable is
uttered in 2d tone.
However, if the speaker insists to pronounce the first syllable in 3d
tone, the next syllable would become 1st-tone syllable. So /jiao3zi1/,
/sao3sao1/ (sister-in-law), /nai3nai1/ (paternal grandma), / lao3lao1 /
(maternal grandma), /jie3jie1/ (elder sister).
Lee Sau Dan
dungping5> Make a correction. Jiao3zi3 is not correct. Usually
dungping5> when two 3d-tone syllables are uttered one after the
dungping5> other, the first syllable is uttered in 2d tone.
dungping5> However, if the speaker insists to pronounce the first
dungping5> syllable in 3d tone, the next syllable would become
dungping5> 1st-tone syllable. So /jiao3zi1/, /sao3sao1/
dungping5> (sister-in-law), /nai3nai1/ (paternal grandma), /
dungping5> lao3lao1 / (maternal grandma), /jie3jie1/ (elder
dungping5> sister).
Those aren't tone 1. They're the neutral tone. Nobody says
<jiao3zi1>. Only <jiao3zi>. Similar for the other examples you
wrote.
Can't you identify the neutral tone?
dung...@yahoo.com
possible and natural, because this sound can be both stressed and not
stressed. When carrying a stress, it is 'zi1'. When speakers of
non-mandarin dialects utter this word in mandarin, they tend to say
jiao3zi1 rather than jiao3zi. There are much fewer light-tone sounds in
the mandarin spoken by non-manderin speakers. This is an issue that
must be testified by the science of acoustics, rahter than assertion.
'Neutral tone' is a confusing term. There is no such an independent
tone category in mandarin. The term of "light tone' is better, as the
sound is just relatively light from the original tone, rather than
becoming 'neutral'. The light tone is substantially and unreasonably
exaggerated by some researchers, and in some dictionaries. Some light
tone sounds are actually tone sandhi. The 'jiao3zi1' is an example.
There are lots of light-tone sounds in my mandarin speech. This is why
I am able to distinguish the tones, and not take your statement for
granted.
Lee Sau Dan
dungping5> It can be both jiao3zi and jiao3zi1. The first-tone
dungping5> zi1 is both possible and natural, because this sound
dungping5> can be both stressed and not stressed. When carrying a
dungping5> stress, it is 'zi1'. When speakers of non-mandarin
dungping5> dialects utter this word in mandarin, they tend to say
dungping5> jiao3zi1 rather than jiao3zi.
And the pronunciation of non-native speakers should be taken to be the
norm?
dungping5> There are much fewer light-tone sounds in the mandarin
dungping5> spoken by non-manderin speakers. This is an issue that
dungping5> must be testified by the science of acoustics, rahter
dungping5> than assertion.
Again, why are you talking about non-native speakers? Should we now
say that Mandarin doesn't distinguish between zh and z, because (most
of) the Taiwanese do not make that distinction?
dungping5> 'Neutral tone' is a confusing term.
But the phenomenon exists.
dungping5> There is no such an independent tone category in
dungping5> mandarin.
Nobody claim it is.
dungping5> The term of "light tone' is better,
I can't see why. There is no such an independent tone category in
Mandarin, either.
dungping5> as the sound is just relatively light from the original
dungping5> tone, rather than becoming 'neutral'.
The vowel involved is usually neutralized. Much like the unstressed
syllables in English.
dungping5> The light tone is substantially and unreasonably
dungping5> exaggerated by some researchers, and in some
dungping5> dictionaries. Some light tone sounds are actually tone
dungping5> sandhi. The 'jiao3zi1' is an example.
A wrong example. Nobody says "jiao3 ZI1". And non-native speakers
are more likely to say "jiao2 zi3", with the tone-sandhi rule applied.
Dylan Sung
"Lee Sau Dan" <dan...@informatik.uni-freiburg.de> wrote in message
news:8764ura...@informatik.uni-freiburg.de...
>>>>> "dungping5" == dungping5 <dung...@yahoo.com> writes:
>
> dungping5> It can be both jiao3zi and jiao3zi1. The first-tone
> dungping5> dictionaries. Some light tone sounds are actually tone
> dungping5> sandhi. The 'jiao3zi1' is an example.
>
>A wrong example. Nobody says "jiao3 ZI1". And non-native speakers
>are more likely to say "jiao2 zi3", with the tone-sandhi rule applied.
jiao3 zi with zi in the first tone which is a high tone sounds odd, affected
even.
Dyl.