Player is loading...
Introduction to Phonetics and Speech
Rob van Son, Amsterdam
Monday, Sept. 24, 2018 · 9:02 a.m. · 44m 08s
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
now
00:00:02
this will be the first talk will be an introduction phonetics and speech we
00:00:09
i asked around we looked at the applicants for the
00:00:12
students for uh the cup was project and
00:00:16
you got the impression that not many of them have had speech of phonetics in the yeah
00:00:22
uh education so we thought before we start and give you all kinds of problems that speech important
00:00:30
matches it might be useful to give you an
00:00:33
overview of the terminology and the principles of
00:00:38
uh speaks like what's normal so that we can contrast with what is
00:00:45
a button logical so i will try to cover if it's not time benetton me and officially ga
00:00:54
physiology out to look at the door phonetics acoustic phonetics and then how you combine these two
00:01:02
what relation and if time allows it sounds say uh some uh
00:01:07
extra about intonation of fundamental frequency so let's start it
00:01:14
this is a a picture of all the organs if the playoff hole in uh
00:01:20
in the uh physics off speech
00:01:23
the brain will because out a extensively later on this are in this uh plane
00:01:29
you found and about the law and the only thing we want to uh
00:01:35
we can say at this moment is it is that saying that buttons yeah
00:01:39
school would be vocal tract to do this would be a vocal fold
00:01:45
and that the hives speech and then we have
00:01:52
as a first important or when the lines
00:01:56
which houses the vocal fall falls the voice box that's
00:02:00
just put this generating before his them mm sound
00:02:05
then we have the old definitely that's the mouth that's
00:02:11
the major park and then we have the nasal cavity ford measles
00:02:16
and uh some of our resonances of cavities and the rest of the hat
00:02:22
that play a whole big i'm damn less important for this call
00:02:29
let's start with voice
00:02:32
this is still a a cross section of the landings it's uh this cartilage
00:02:38
uh some of here and if you look inside it you say to
00:02:43
fall surf to flaps you said a vocal chords are and
00:02:51
the um muscles attached that can close them open open
00:02:58
and somewhat more
00:03:01
detail
00:03:04
what happens is this is the way you see y'all they
00:03:10
that's the like this they move like this is what this time too
00:03:18
show here and i have us movie so what movie about it
00:03:26
yeah several several uh
00:03:30
states
00:03:32
you can uh
00:03:35
why don't them for inhalation
00:03:43
if you met with them you get this thing you don't hear a voice but you hear a whispering sound
00:03:53
and if you want to for nate
00:03:56
you bring it together and you press yeah through them and they work like
00:04:01
like when you do the ritual that the same principle works here
00:04:08
and by the stretching or loosening them you can uh
00:04:14
manipulate uh the frequency of five of it
00:04:22
and no man has generally have long well folds vocal
00:04:26
cords can women and that translates in that
00:04:30
man have a low voice and women that's what government does the the pitch of the force
00:04:39
uh_huh
00:04:41
next that this is just a mm voice a sound of that force
00:04:47
now we want to make speak speech but it so all we have a lot a number of
00:04:53
organs that can be used to manipulate the airflow and serve and that way the speech
00:05:01
uh and the so called the articulators the most important ones
00:05:05
hardly draw you can open and close your mouth
00:05:10
then we have to lips which you can last together
00:05:14
or open or lucerne was the or you
00:05:19
can do yeah it's colder rounding you pursue lips the most important one must be long
00:05:29
but such an important organ that many languages language itself
00:05:34
how ah in the name of language what language is actually it was football
00:05:42
i don't listen very complex or articulate on i forgot
00:05:47
the teeth but there are harder a fixed
00:05:51
it's very complicated or when and if you look at the movements it has
00:05:57
three or four independent parts market can move independently
00:06:03
the one that's most mobile is the tape you can move the tip of
00:06:06
the tongue independent of the rest of the topic so you can
00:06:10
put the tape of the tong again say this rage uh of aerial image and
00:06:16
now even if you bring it on back formed or high or low
00:06:22
then we have what's called the body of the tong or the back
00:06:27
um and this is sometimes divided into two or into three i
00:06:33
think we do can high be put high or low and
00:06:40
most of 'em clear examples d. e. sound where you put
00:06:46
it on the back of the tongue against that
00:06:50
then we have the wood of the tong and it's not used to that many language i believe which you can
00:07:00
pressed against the back of the yeah uh of the findings of the fall
00:07:06
and the most clear for sounds it's the it's it's not have like the the the the
00:07:13
cripples sounds of that rabbit where you put your will to sit on the back
00:07:20
then we have to feel an uh this is the all cavities than most sounds or the
00:07:25
you use all kathy but sometimes you let yeah go through the nasal nasal cavity
00:07:32
and that you do by lowering the loom this the last articulate i will
00:07:38
um
00:07:39
talk about um some time to do it but files so that fans you have
00:07:44
nice eyes files and and there's ice files in many languages have that
00:07:48
and that's that's you achieve by lowering or high uh the feel him an opening in these all yeah
00:07:55
you hear it you a clear difference in the sound but uh yeah it goes to the nasal cavity one mall
00:08:05
so that's the
00:08:07
say the anatomy all speech
00:08:13
so then let's give an example of the most simple class of this sounds vowels
00:08:23
um because they x. play they they gave them a slate lot
00:08:27
of the yeah the the principles of speech off um
00:08:33
all files are articulated with a mouse open so there's always an open connection
00:08:40
between the vocal cords and he and the uh the i'll go yeah
00:08:47
what's different between the files that's the position of the tall but don't can be
00:08:52
good the tone can be to the front and the high like an e.
00:08:59
can be to look back and low like oh you can open your mouth in ah
00:09:07
your some more clothes and then the others in between
00:09:12
now that these are the the three three extreme mouse like he is the
00:09:18
most high high uh uh you know almost close under don't most long
00:09:25
who is don't most back wrote ah is maximally open
00:09:30
mouth i over sob to we like if your do nothing
00:09:36
you relax completely and you let your mouth completely relaxed open like a get a
00:09:44
that's the that's what i found was that uh sound is that easy
00:09:48
sound to make because you actually do nothing everything is relaxed
00:09:54
but if you know
00:09:56
you can call it steps form a beep when the one and two one
00:10:02
and a and a ah that's uh uh i couldn't uh uh
00:10:09
i don't see is of files where you just open your
00:10:14
mouth and put your tongue at actually open house will
00:10:19
from the whole then just do the same from the pool
00:10:23
ooh which also surrounded you also pursue lips you go for
00:10:28
two two uh all i all all our ah
00:10:35
that's like another c. a.s and then we can go from a while back
00:10:40
about and that's on sounds yeah in that you have to you
00:10:46
which is in between somewhere close to do we
00:10:50
going from one to you so that's the
00:10:55
that's and then you see this is the file triangle that's the way the onus of also say if you
00:11:01
go from e. tool to i we can make this list can make this list so we fill this
00:11:08
triangle and whatever language we go we can put the vials of that language somewhere
00:11:14
in here based just on the position in its uh don't fold or back
00:11:20
or is the mouse close open are the lips pursed on mark still third dimension
00:11:27
and the languages that only three files that only had the one ah
00:11:33
damn any language that have only five files like japanese or spanish
00:11:37
and that is generally the u. i. and then an ad
00:11:42
sigh e. a. x. i like sound and then all was like sound
00:11:47
and then yeah with progressively more files like the
00:11:50
chest feldman of tongue files and in uses
00:11:55
eighteen well somebody if many more somebody but at the
00:12:03
at this at the moment the gist of distilled and you get
00:12:06
extra uh over the fact that in on discuss now so
00:12:13
these are just the vowels obviously we have a more also have consonants also some more complex
00:12:22
and this is a list or else did the eva eva i'll so bad the phonetic alphabet them
00:12:29
all well almost all the continents of the most common consonants and they are divided in
00:12:37
by but i yeah voiced or not is he is or this
00:12:43
uh but uh they ah a howdy i made and where we put out on that
00:12:50
like in the in the uh without so let's first look at the voice
00:12:59
question is in the states we can make a sound with the vocal folds five i think
00:13:06
that's a voiced sound or the vocal song falls not vibrating that's an invoice so
00:13:12
well we have to it's we have is is you have a
00:13:16
better which is forest not but it is not voice
00:13:23
uh this is what we see in the recording in the bay
00:13:25
form of the recording she to noise and some kind of
00:13:32
it's a regular vibration is you some kind of regular vibration
00:13:36
in the closer yeah and you see nothing yeah
00:13:41
so let's look at the next to mention the man of creation
00:13:47
that is a constant is basically a sound that you make with the construction in your
00:13:54
in your vocal tract somewhere you block the vocal tract or almost completely or complete
00:14:01
and the man of articulation this is how you do that
00:14:08
the degree of construction about on not to generate a a fricative noise so i'll
00:14:17
we can make a in in all of impulse in increasing the construction
00:14:23
because start with the uh vowels which are open vocal tract
00:14:28
we have diff don't it's a a kind of files but then to one oh
00:14:34
we have things illegals semi false yeah like files
00:14:39
but they almost closed vocal tract like
00:14:42
yes yes yes well while or being well yes
00:14:48
all these cases you you close your vocal tract almost
00:14:54
uh uh but not why and you hear the dip
00:14:57
that we have the closest where the block completely blocked the vocal tract like that
00:15:04
and then release it you we're the this to stop but has
00:15:08
no sound nothing and then you hear the sound of release
00:15:15
then we have nasal swear we broke the vocal tracts but open in the the the nasal
00:15:23
if you live so that yeah can escape would not
00:15:26
an uh do not like an name on mm
00:15:33
and the i think it s. and that we we make it to the noise the
00:15:37
press the ass who really know construction and you hear like us off sound
00:15:46
that's the basic building blocks of speech you make speech but alternating this kind of stuff
00:15:56
but that does as the next to mention the last dimension i will i will handle is
00:16:01
well do i make the block in the vocal tract that matters like
00:16:06
involves it met up mattress where yeah i put my tongue
00:16:10
in concert that met that's why i put this block which is this construction
00:16:17
and there are many places to make it big construction from front
00:16:21
to back yes we can put the lips together like a
00:16:26
we can put a lip against the that he liked off
00:16:33
or we can put it on at the teeth into dental is like in the i mac and playing like
00:16:43
you can put it on the tip of dongle like just behind the bees that's the
00:16:48
most frequent the most easy to make a constricting that's the d. and d.
00:16:54
and then you can put the back of the tone against it but let's let's it's like a nice
00:17:03
or much set them back like the in famous that uh when you put
00:17:09
you make the the construction like here and some languages uh many languages allow you to
00:17:17
put to put off the tongue against the back in that attic it's quite good
00:17:22
speeches and the last one is you can make a construction here like and
00:17:28
h. sound they use the focal uh the vocal folds to generate movies
00:17:37
these are the places of articulation that all with men of articulation in place of articulation and
00:17:42
voicing you can identify most almost all full means in speech so if you know
00:17:50
in a certain place of speech the speech according to me the place of articulation man of articulation
00:17:56
and the voice saying you can almost always identify the phoneme that has the would use
00:18:05
so
00:18:08
the next question is designed this was a talk about i'd isolated full names but for
00:18:14
the entire not use isolated they are used in combination what combinations i use
00:18:21
in that first day
00:18:25
for every will i tell you that will be an exception some language so let's forget that let's just take the
00:18:32
what's common in languages and in common languages is that
00:18:35
languages are organised in syllables speeches organise and syllables
00:18:39
what's a syllable syllable is a nucleus and something around
00:18:46
the nucleus is there's always a linked list and the new places i doubt
00:18:51
vile all this stuff is either a ah i. r. e. d.
00:18:58
can any valve with this uh well
00:19:02
and it the articles and surrounded of constant before our whole onset constant after what are called the coda
00:19:11
uh and what we call a time when you are harming you combine
00:19:17
the new to this with the the nucleus with the coda
00:19:21
then that's the the the the normal time and go that is just any combination of sounds
00:19:28
a consonant an onset too so if i say plant p. l. the pool is the onset
00:19:36
yeah ah yeah plan as the the new clothes and then it is to go yeah
00:19:44
every language has distinctions in how constants can be
00:19:48
combined to form an onset or coda
00:19:54
and the out double so different in every language the awesome make our rules that tell you what is allowed
00:20:00
or not but i will not discuss that but just to know i can say plant helpless line
00:20:08
but i cannot just makes these i cannot say in nepal that's
00:20:14
not coming is what you know or loop on gas or
00:20:21
the holes that got from which sounds can be combined
00:20:26
in a also gonna kodak will fall back x.
00:20:29
so mine languages don't have go dance like japanese on shoe code as
00:20:36
japanese has add a a new colours or a consonant mucus
00:20:43
of a beef japanese well doesn't even our continent continent clusters
00:20:48
not doesn't allow constant doesn't so a backpack that is it possible
00:20:54
a japanese work but uh all my name is not
00:20:58
a possible japanese work because you can't have
00:21:03
um p. after all comes close it after only plus i fall in japanese
00:21:10
but it's english allows clusters that can be quite cool that's just a to f.
00:21:19
and one point of to buy you doing uh in phonetics on one language this
00:21:25
is one of the things to try to find out what can be comeback
00:21:31
so
00:21:33
now let's stop works i've got or phonetics and go to acoustics phonetics
00:21:42
we talked about what people how people used yeah articulate of now
00:21:48
it was look what this what are the sounds we see
00:21:53
in uh well we here come on top of all the sounds just files
00:21:59
are best understood and i will talk about them or some more
00:22:03
um because files and formal lots even uniform and uh
00:22:07
continues class not only inoculation would also in analysis
00:22:12
mm deposits of problems are rather complicated at the clock of continent has
00:22:16
its own peculiarities that um make it difficult to to channel ice
00:22:22
but um for a stock how do we study uh the acoustic phonetics that is by using the
00:22:31
spectrum if you look here is this survey form of an ah i would give the sound
00:22:39
it's my ah
00:22:41
ah so this is on the stain ah
00:22:46
you see a nice way for
00:22:48
i'm not very informative if you make a hundred milliseconds you say
00:22:55
you get this this would be getting a pattern and every
00:23:02
after the weight is one vocal fold vibration so at this point the fog
00:23:09
vocal fold open yeah capital and uh then at that point close again
00:23:18
and then it goes open again and close to get the that's the the speed with which the vocal
00:23:25
folds flat cash into each other at closer determines the loudness the perceived loudness of the song sound
00:23:33
um and the
00:23:35
uh the frequency with which it a really flap ready
00:23:40
the friday is the page e. uh_huh now
00:23:46
using not only that this repeating high a way for which
00:23:51
also still kind of all vibrations inside each day
00:23:57
and we can look at the frequency of each of these little vibrations
00:24:03
the big once a little once that is a spectral analysis i you
00:24:07
can all turn the frequency of the of these vibrations against yeah
00:24:11
uh yeah yeah and the two or our and then you get this kind of
00:24:17
that is i think we take this one and we calculate for each of these vibrations uh a
00:24:23
difficulty in the power we get this kind of that and i'm like you always see is
00:24:27
this is the page this one the first one is the peaches at the free fundamental frequency
00:24:33
and in multiples of this frequency you see the speech
00:24:39
that's so these are the the overtones the fundamental frequency any overt don't
00:24:46
and this works this like a stinky marker but you probably haven't physics class
00:24:52
how's things vibrate today for bait with the different
00:24:56
frequencies that are multiples of each all of
00:24:59
the fundamental frequency and that's what you see here what you'll see is that that is
00:25:04
not all are equally high powers not evenly distribute it
00:25:09
and we see if we do what we can do a different analysis that
00:25:13
we can get then the to see that uh two bumps here
00:25:17
these are important designed of older formant frequencies they will discuss them
00:25:23
next they want also want to show is this is the pitch is different about this the
00:25:28
frequency of the first peek over time you see i i make a fairly stable
00:25:35
uh uh uh for is a very stable for years and years and years and we stayed voice
00:25:41
yeah the the ah apologies but this is not the people i'm not able to
00:25:46
either funny at all or keep a stable uh frequency not just
00:25:54
looks all kind very stable and and repeated but if he
00:25:59
do this analysis which was of a source that overall the whole of this
00:26:05
four seconds it believe this oh three seconds what is it two seconds
00:26:10
you get you see that keep each other yeah but i bought that that's because
00:26:14
i'm not quite stable there is some variation and it's add some more studies
00:26:21
so this is what you see in the spectrogram like the spectrogram i show
00:26:26
that before ah it's quite simple but it's desert when they say ah
00:26:34
that has no that's aspirations so i'm not very good at it ah ah
00:26:40
i'll ah but this is what you get if you get the more complex sound is the vowel
00:26:48
it's what you saw with the and this looks quite like ah with two of the bands here on the formants
00:26:54
and this is again and ah and the bans user to do our informants again you see the petitions
00:27:03
and in between you shaded the closure here it's just noise
00:27:08
then you see the release because really short because
00:27:14
oh
00:27:15
and then the aspiration here isn't ah like sound this is quite
00:27:23
more mole a requirement in normally see
00:27:27
now see these formants what i show chill was that you had these buttons yeah now we
00:27:33
can do nice analysis to with all the files and we see that these buttons
00:27:38
right so if we look at the e. e. for instance you have a vase low been very low fit this
00:27:44
is frequency again from zero to think you know had missus powell get and for easy if we go
00:27:52
to the left for through the ah what we had before basically two peaks
00:27:56
of the am is also well is that these two p. h. one
00:28:02
quite high six from not helps like or higher and the second one is eleven twelve one of her life like that
00:28:10
if you go to be easy that the first week is much lower well t. v. or not hurt
00:28:16
and the second rate is very high like over to kill half of
00:28:21
them and there's a third week even higher but that's normal
00:28:26
if you look at the new whereas this this this test line you see that this first formant
00:28:34
is like the same decision l. c. and this gives an indication what this first formant means
00:28:42
basis mouth closed his mouth open so the first formant gives you an idea how far the mouse is open
00:28:51
but the second formant of two who say well it's like seven
00:28:56
home about it's all it's just a little higher than
00:28:59
that ah often it's quite close to the first formant of the ah the second formant of that was quite low
00:29:08
the third we ignore yeah because it's not not that important for vowels
00:29:15
now what's the difference between in which the position of the top this means don't hang don't well sort
00:29:22
of took the position of the second formant tells you something about that on this in the market
00:29:30
i simplify here i see some people this i'm simplifying yeah but
00:29:37
nice you plots the first the second formant all vowels as we've gone for i believe this was
00:29:45
uh oh this is a i mean i can files you get this this plot the plot easy get them here is the
00:29:54
first formant second formant we get them here the plot was to get them here a lot ah i mean i can
00:30:01
has no ah but i like sounds all here hi everyone
00:30:07
intermediate f. two and all the other files like a a a a a a
00:30:13
ah or who that's the same order as in you well triangle applauded here
00:30:21
for historically since they do everything to confuse do by time
00:30:28
my uh how taking and and a mailing but it's the same or like a a a a
00:30:35
an ah a are all that's this this although
00:30:42
it this order is this or no
00:30:46
any in between
00:30:48
and these
00:30:51
so what is seen this block is that vows have not one
00:30:57
the size of one of two but and yeah yeah easy d. v. d.s um overalls
00:31:06
this overlap in running speech this is quite well articulated when it's busy overlap
00:31:12
and the articulation maps to d. f. one f. two i'm back so you can given
00:31:18
a certain articulation you can predict where the f. one and s. two will be
00:31:23
and if you have to have one of two you can make a guess about appalachian what i didn't tell
00:31:30
the one complication is using let's if you pursue let's knock this should will change the first formant
00:31:38
and so you can have uh you can have this the same position in this have one of two
00:31:45
um uh uh yeah yeah a space
00:31:51
using the same position of your song a different position the tong but by compensating for your lips
00:31:58
does a lot of those lot of uh that's that's kind of a but then see
00:32:05
um now this tongs also this domain high on our on all sound
00:32:11
it's simply of all the to target the two positions
00:32:14
you started certain point and you go somewhere else
00:32:18
in this i've one as to plot it turned it around them leave again
00:32:22
but you starts on that you go somewhere you start at in
00:32:26
our is you started off i'm going to do that 'cause i'm gonna not of time so i will want to do some of those
00:32:37
now how can combine this idea off
00:32:41
formants from acoustics and articulation
00:32:45
formal and the comics and they're not gonna go phonetic how can we combine them
00:32:49
how can reduce that now this is combined in the source filter what l.
00:32:55
um speech sound comes file you have a source of the sound vocal fold vibration
00:33:03
you have a vocal pick shape could use that acts as a filter
00:33:06
and you have the movement of the articulators that manipulate the filter
00:33:11
and we can infer what i'm not to me and fish elegy from bay
00:33:14
format spectrogram and we can uh uh info did sound from the anatomy
00:33:21
and it works like this you start on the left of it the
00:33:25
yeah floats to block this last friday so the yeah the the
00:33:30
do l. speed goes up and down up down up down it's faith that into
00:33:34
the vocal track and what comes out i find patients that i filled out
00:33:41
if you work in the spectral domain if it's a spectrum of bases all these uh
00:33:46
also don't the fundamental tall but only also don't also wasn't that's
00:33:51
what the soul before then you have youth characteristics all see
00:33:57
the vocal tract has the speech the speech formant peaks you said the formants
00:34:03
it s. earnings at certain points and it and it has an it's like it it in a it inhibits although
00:34:10
frequencies some frequencies can get through all those not so you multiply these
00:34:15
two is the frequency access and this is what comes out
00:34:20
so knowing before the the the fill the characteristics of the vocal tract you can
00:34:27
the the x. out the sound will
00:34:31
look the spectrum of the south and how little sound for the input
00:34:36
which means that if the input it's damage out you will see the output you can put the what can
00:34:42
the output will be cool well and if this is the fact you can also see what's going on
00:34:50
this again uh an x. for the noble example
00:34:54
you have a source you have to fill the characteristics of the spectrum of the source spectrum of the filter characteristics
00:35:01
filters in the provoke like this is something that happens at the lip gonna
00:35:05
ignore that it to boost the high frequencies this with the get out
00:35:10
a four man that a low for a should get this kind of spectrum
00:35:14
this is the output informal menu get it's kind of spectrum which sparse so
00:35:20
a bit because the public would come
00:35:27
can we do something but that yeah we can make we can put these fill the characteristic by loosing
00:35:33
is it i met i of x. say a photograph of the uh of the all time
00:35:40
you measure which the only thing you need is to do to enter yeah at each point in this
00:35:48
in this vocal tract so you make a a pass from the clock this on the vocal fold court
00:35:55
full mouth weekends and at each point you'll match your yeah yeah
00:36:02
this is the i function this is handled i i copied it from somewhere is handled a little bit
00:36:09
uh or cork but this is the any offence you you say you go from
00:36:14
from here to yeah and then at at youth at each point you tell us what is the ad
00:36:24
then there's a simple no simple doesn't mathematical click to go from here to use all
00:36:31
the characteristics show with each of these yeah yeah function should get fill the characteristics
00:36:37
so for but this plot we can predict what the sound will
00:36:43
look like or how little so
00:36:47
and this is such an uh if any of function
00:36:52
it's much too detailed we don't need all these detail detail you
00:36:56
can just use one set of troops this is the
00:37:00
this is one of the most important analysis uh uh within a speech
00:37:04
in of acoustic phonetics how to get from speech to this
00:37:12
and we had some at a way to record the sound
00:37:17
and determine this without having to go through the and not them it works both ways
00:37:27
and to convince you that this works i just
00:37:31
but do some of these these are these these elephants and think of this
00:37:35
as a as a fruit like a float around and this afloat
00:37:41
and uh the food is i made two may to sound like ah e. and who
00:37:47
and i think in this sound this is the uh uh an artificial sound sort mm
00:37:53
um this is what the vocal folds sound like if you wouldn't have a vocal track
00:38:00
and this is what the flute says if oh oh this
00:38:05
is simply that brought a sound fat into this shape
00:38:11
mm mm
00:38:15
who who does it contains different at this phonetics
00:38:22
of locations that this works this model
00:38:27
this is the way in what this is
00:38:31
well that's normally goes
00:38:34
and someone for educational purposes a japanese just the i make these
00:38:40
these fruit like thinks it would on plastic i think
00:38:46
with these dimensions just to generate the phi files of japanese
00:38:51
they work and i give you the sound here
00:38:57
ah
00:38:59
ah it's still it still sounded it to this watch this stuff each
00:39:11
ah
00:39:14
ah
00:39:16
oh
00:39:18
oh
00:39:21
i lost all equally well
00:39:27
so
00:39:29
how far can you go in this model this is one where you also
00:39:33
include do nasal cavity a and the uh other cavities in your hand
00:39:40
and to show you how very far you can go i'll
00:39:44
show you a some one person be the buckles
00:39:48
has developed this into an articulate was it this is this is the uh say the vocal tract
00:39:57
this one and he moves the tongue and lips to do this
00:40:03
that's what i noticed when people
00:40:06
that's what i noticed one of such people
00:40:11
i have the this is surely generated like sound a
00:40:15
force of four source on in a speech out
00:40:20
to be totally honest this is not working very well for the
00:40:24
fact that this and some a consonant sounds so he
00:40:29
cut some corners at the concert sounds but the valve and the dates involves all uh
00:40:37
a realistic no i well i'm almost out of time so i
00:40:41
will handle that one thing more what's important escort creation
00:40:47
um if you say e. d. r. that or will you will see
00:40:53
that the this uh uh places of attention x. a photograph
00:40:58
you see that the tong it for the d. at the the point the tang shapes are different
00:41:04
its shape of the e. ah and one i retained but
00:41:09
only did don't tip is moved to hear this difference
00:41:14
in this sounds into the sound of the d. and then the on and also
00:41:18
and this is called code relation you don't do everything you need to do
00:41:24
if we we take a full means uh started and you have to to move
00:41:28
your tongue from pocket want to target to to target c. into target for
00:41:33
in a complex sequence you don't will in reality use
00:41:38
the least effort to get to be that
00:41:40
is necessary and it will not hit the intermediate targets and this is called at relation
00:41:48
and uh this is you can measure that that is easy to find what's important for speech pathology is
00:41:56
when time in the time of these uh movement as extremely important
00:42:02
if there's something wrong with the timing of these uh these
00:42:06
uh uh transitions the these audible and they can complicate
00:42:12
your speech and i think i stop here because i want to pay
00:42:16
for the next speaker whom for what he said suppose uh
00:42:24
his computer i will answer questions
00:42:29
yeah
00:42:35
if
00:42:37
okay listen if you have any questions that i
00:42:42
can answer that out of other pictures
00:42:48
yeah but
00:42:52
yeah
Conference Program
Introduction to Phonetics and Speech
Rob van Son, Amsterdam
Sept. 24, 2018 · 9:02 a.m.
297 views
Dysarthria
Marc de Bodt, Antwerp
Sept. 24, 2018 · 9:45 a.m.
303 views
Children’s speech: development, pathologies and processing
Alberto Abad, Lisbon
Sept. 24, 2018 · 2 p.m.
102 views
Speech after Treatment for Head and Neck Cancers
Michiel van den Brekel, Amsterdam
Sept. 24, 2018 · 2:45 p.m.
144 views
Speech therapy
Marc de Bodt, Antwerp
Sept. 25, 2018 · 11 a.m.
114 views
Assessment in speech disorders
Virginie Woisard, Toulouse
Sept. 25, 2018 · 2:45 p.m.
140 views
Recommended talks
ESR04 : Acoustic speech based monitoring for telehealth and predictive analytics
V. Srikanth Nallanthighal
Sept. 4, 2019 · 2:54 p.m.
160 views