Player is loading...

Embed

Copy embed code

Transcriptions

Note: this content has been automatically generated.
00:00:03
together was rubber with organising everything today so his name is on here
00:00:08
and the actress cause lies and then stop go up one was and is doing the
00:00:13
project for topless in amsterdam from angry and we have a corporation university of and
00:00:20
yeah some other uh universes well so what i'm gonna try to tell you i'm at a medical i'm a doctor so
00:00:26
actually patience and um so we also cooperate with engineers and
00:00:31
uh foundations and uh as a piece to uh
00:00:35
to be able to um and then something and uh about uh about the voice
00:00:40
and um and speech of patience and weekly basis would have the neck
00:00:44
cancer so i'm gonna give you a little bit overview what is had a
00:00:47
neck cancer cancer of the mucous membranes of the mouth and the throat
00:00:52
and the larynx and what will happen to these patients from between them give you some examples and we also working on them
00:01:00
how to make it automatic a speaker recognition and trying to uh do that
00:01:04
together with um and rip and and the kent university so um
00:01:10
well try to give you an overview so had a neck answer most of you will well not know much
00:01:15
about it probably with it's cancelled it originates from the time for example or the throat or the larynx
00:01:21
and uh if we treat those cancers and it has influence on the
00:01:26
speech um well what how does it how does it uh
00:01:30
ochre his head and neck and so it's quite a rare cancer if you compare it to breast cancer or colon cancer
00:01:35
um and it's about um in the heavens about three thousand patients
00:01:41
and so uh we've seventeen million people in uh the handlers
00:01:46
but for example in the south of europe uh in spain and portugal is much
00:01:50
more common it's about the double incidence yeah because of the smoke more
00:01:54
so i come to the incidence and uh and the reasons why you get it so
00:01:59
i'll go and smoking on the big things that cause this yeah and this cancer so what you can see
00:02:05
happening now for example in united states was smoking we went down is that the the incidence is decreasing
00:02:12
uh but that's not so much the case yet in europe because we still drink a lot and we also still small
00:02:19
mainly in older patients starting from fifty some countries in in
00:02:23
your behalf snuff tobacco settings stuff like in sweden
00:02:27
and the and the norwegian countries beetle not is not so common in europe of course was more india
00:02:33
h. p. v. virus is really coming up we know the h. p. v. virus from cervical cancer in women
00:02:39
um and it's coming up as a cause of throat cancer at the moment more than half of our
00:02:43
throat cancer caused by h. p. v. which makes a sexually transmitted disease
00:02:49
uh in d. v. major fairies cancer meaning china and india also
00:02:54
uh but also predisposition plays roll so if your father or mother has has a cancer
00:02:59
lung cancer skin cancer your risk is about five times higher also to get
00:03:05
certainly some genetic things about and also died and i'm not uh so many things about diets i'm not gonna go into that uh
00:03:12
um you see this is the incidence in in allen's uh there's very comparable to the rest of europe so we still see
00:03:19
so now we are uh are uh around here around three thousand patients that is rising still a little bit
00:03:25
and the rise is uh i'm not so much because it we we smoke
00:03:29
more or drink more it's more because the a the population is easy
00:03:33
and so we see that you know in all your work but uh in average we get a um everybody's getting older
00:03:42
um easy incidence of larry's cancer in in all different country you're
00:03:47
so you see the new bars are the incidence of it is is very different in
00:03:51
all countries depending very much on your habits of uh smoking and drinking um
00:03:58
and a total european is here but you see these countries have
00:04:01
a much higher incidence like reese and a romanian bulgarian angry
00:04:07
very high incidence and also the mortality or mortality is the black bar
00:04:12
and it differs quite a lot for example these two countries today is is about
00:04:16
the same with mortality in a miller is is much higher than in luxembourg
00:04:21
and it has two courses in fact yeah i'm one
00:04:25
is uh they uh are found in earlier stage
00:04:29
more in general in the more richer countries uh this is
00:04:33
found earlier because patients go more easy to doctor
00:04:37
and that also the quality of the care is is probably a better in
00:04:40
some countries but you can see the very big difference in europe
00:04:47
now if we want if we are treating these pages we have of as as doctors and
00:04:51
peace and uh we have a few agendas of course we want to do a few things we want to cure the patient
00:04:57
but also we are looking at mobility because well we can uh if you have done cancer you
00:05:02
can cut out the whole time of course and the patient will always have good margins
00:05:07
um but then of course speech is terrible so there's always a compromise between
00:05:11
mobility what we what what we what we do to the patients and did you right so
00:05:16
in there is cancer the best here is to take out the whole there is it all patients
00:05:20
but of course you don't wanna do that's innovation because then the mobility is too high
00:05:26
uh so long term quality of life mobility play a big role at the bottom and still the cause
00:05:33
i'm not the biggest role but uh you can see that coming up of course
00:05:36
everybody thinking about because because it's a health care is really costing of things
00:05:41
and patient preferences are uh in the past uh this was not
00:05:45
so important i think the doctor decided that more and
00:05:47
more i think the patient is deciding on the street and you can um have a say in that as well
00:05:55
i know what mobility do because i'm treating these patients but you can imagine that swallowing and chewing if you
00:06:02
three things you had a big influence but also speaking
00:06:06
articulation and voicing really can be effectively nice page
00:06:10
apart from those two things uh mainly we're now today i mean it over but speaking voice
00:06:15
we also have a static body image a taste smell feeling social impact
00:06:21
and on work so um these the having these you was is not really fun
00:06:26
i'm gonna show you a also some pictures of treatments so if if you don't like
00:06:31
lot or a terrible things you wanna do the other way um but um
00:06:37
sometimes it more cryptic lack and this is a symbol for surgery and these are just butchers um
00:06:44
so what are the treatment options we have surgery you can do that with a knife but you can also do there's a laser always
00:06:50
a p. t. with the robot eventually even example where your therapy was on mine is chemotherapy
00:06:55
very often we give were you so p. plus chemotherapy and all kind of communications
00:07:01
well these are some other things got immune is there be probably people heard about
00:07:05
is really coming up as an important modality at the moment very expensive still
00:07:10
but certainly a um for the future so some examples of dealing with these other vocal chords
00:07:16
and they should be nice and useful but you can see it is a small two more vocal chord
00:07:21
and you can feed is very small to most you can just you can radiate them
00:07:25
they very good but you can also get them away with a laser for example
00:07:29
or in some countries where you don't have a good a good places system they also just open neck and they do it from the outside
00:07:35
but uh that has a big uh mobility of course so small to most don't
00:07:39
have a big impact on quality of life maybe a little bit of worsens
00:07:43
but not much them more than that is a ten cancer so you see how you put away part of
00:07:48
the time has not begin implications for the function uh we can we take out parts of the term
00:07:54
and patients really have a enormous uh a possibility to adjust
00:07:59
to compensate for that although you can here it's uh and this morning i did a cleaning and read some pages part of them
00:08:06
and you can always hear it of course because some sounds are some articulation it becomes difficult
00:08:11
because the town is a bit fixed in the mouth but um but not so much
00:08:15
it becomes bigger this page more than half of the time was taken away
00:08:19
and you can see it was a skin flap stitched in place you can take
00:08:23
the skin flaps from all kind of places but you can imagine that um
00:08:27
this part is not really moving vacant reagan's would have difficulty talking with this time
00:08:33
but also that in principle they will compensate partly protests but it's not always easy to um
00:08:41
and to counsel the patient to really tell what's gonna happen with you know bigger operations also possible
00:08:48
like taking away part of the mental you we put in mystic piece of people off
00:08:52
on the lower leg can see your well taking away half of the mouth and
00:08:58
if everything fails you can get this and terrible things that patients lose their memorable and they haven't known for
00:09:04
total or cripple cannot eat very difficult speaking and it's
00:09:09
all these things can also go wrong unfortunately
00:09:13
this page no hold on it to be removed from normal term and the only big flap in the mouth you can
00:09:19
imagine this lab is not really moving i'm gonna show you some examples allow these patients can still speak a um
00:09:27
if you uh uh examples i'm gonna show you this is a ah how you do there's a robot
00:09:32
well it's very fashionable nowadays to operate with a robot um
00:09:37
and and the this is the robot you can see the very big machines still in the future i hope we
00:09:41
will get some almost because they are effective very big and you have these things going into the mouse
00:09:48
and you and steered the robot with uh with your hands um but it does it so that
00:09:54
you can see how this is it you more here at the base of the town
00:09:58
and uh and so you you just go to the way this is yeah an electric online
00:10:03
um
00:10:05
it's still a distance and the advantage over white what do you do the robot because you can we can small hole
00:10:11
and still have movement and you can work on the right angle because you work with the camera
00:10:16
that's um that can can be at thirty degrees or forty degrees of fifty degree
00:10:20
so you you have more possibilities to operate a round angle and um
00:10:27
so in that way you can do your surgery um
00:10:32
this is a graph to show you a little bit for publication and that the
00:10:36
place where you could away part of the time has influence on the voice
00:10:41
so this was the intelligibility intelligibility score from patients with the base of time cancer
00:10:46
and oral talk and so you know that the time has an oral part and based in part because you can base it on
00:10:52
it how do you have any influence on intelligibility you can we take we
00:10:57
chance of a base of them um but in the world on it becomes much more important
00:11:05
so since a couple of years what we do in all these patients uh
00:11:08
our as a piece to a like a class guess experience with that
00:11:12
hey we send them to be as a bee in four before the operation
00:11:15
and that time interval after the operation so we have an idea what
00:11:19
the patient has before treatment after treatment and have to fill in all
00:11:23
kind of things or kind of scores um uh and we makes
00:11:29
mm recordings of the of the of the voice of course and them
00:11:33
so it's it's a nice way to do research in this field
00:11:37
i'm gonna show you few examples a patient's um
00:11:42
um we had been operators in all kinds of so i took some examples of
00:11:46
places who have some complains of course so the most the biggest surgeries
00:11:51
um yeah you know me better you buy ice in this but um
00:11:57
so this page was eight years after surgery and well yeah you look so he's not really able to be probably
00:12:06
uh_huh um how do you deal with it and it will uh huh
00:12:11
uh_huh leave it and go home and we cannot really sorely has back through
00:12:17
but who uh who are the about hogan then you you mean uh_huh
00:12:22
mm maybe no yeah i would be well one oz
00:12:27
digital the new shows no i don't know null on lose line it you
00:12:33
will allow you was already here well at least maybe i just
00:12:40
a non sound and you go oh oh yeah he'll let alone and not you
00:12:46
know so there's a lot of not announce the so one uh huh
00:12:51
ooh ah all all alone on well why don't you
00:12:58
maybe difficult and i also something example there's another patient mm yeah and um we could you know uh_huh
00:13:05
uh_huh okay so they very often try to use the looks more you can see uh_huh uh_huh uh_huh
00:13:12
uh_huh and ah ah but you didn't put me off uh_huh uh_huh uh_huh uh_huh
00:13:21
uh_huh agreements we don't ah ah yeah are you lonesome huh
00:13:26
uh_huh yeah i hope you get the impression all those that each utterance also some
00:13:31
english recordings isn't english recording maybe nice ocean or not should months post ah
00:13:38
ah my multiples huh i mean uh_huh
00:13:48
industries remove one mike chow who are slightly more
00:13:53
what's happened why sharon so much as
00:13:56
removes place um especially um farmer for consumers
00:14:02
show actually an agent johnson tendon
00:14:06
enter ah yes was going over down to um our
00:14:13
what you wanted and here i am now my chop it in its very intelligible i think
00:14:18
but uh they are very often compensate for kind of things and as appease play a
00:14:22
big role in that of course to teach them how to do this um
00:14:29
well these are some because it's in these may be nice for you to this and is so these
00:14:33
pages really had a total go check to me yeah yeah yeah yeah yeah yeah yeah yeah
00:14:41
e. e. n. e. i. e. e. e. bay
00:14:51
yeah you you you
00:14:54
yeah
00:14:56
again
00:14:58
so are well i think you handicaps can be really quite big um this is a patient uh of mine uh
00:15:06
also the total bill sec to me she was quite young at that time is probably about eight years before disk in a film was made
00:15:14
and um so so basically we have an enormous energy to get out of
00:15:18
this you she changed professions you became a vision therapist of racing horses
00:15:23
um which of course you don't have to talk to the host too much and
00:15:28
but um she became so good um and she's that she swallowed again should not time and
00:15:33
she even gives less is not to um paper on people so well quite incredible
00:15:42
our would believe you you use yeah you did they use humour yes
00:15:47
you do you view and so she sure was firstly risky
00:15:53
recurrence we had to take out the hold on the to again and you will you know that was a
00:15:59
you know a little apples from all all the use
00:16:03
like a yeah it's real i would establish a
00:16:12
oh oh oh
00:16:18
oh oh and ah ah yeah ah and well you might eighty
00:16:28
oh oh oh oh especially speech very well i'm sure
00:16:35
the other one without an oh yeah oh
00:16:42
a man so patients to develop their own strategies but you of course also the as a piece
00:16:47
can help them and uh we hope also in the future all these automatic here um
00:16:52
uh assessment techniques will give more direct it's a
00:16:56
fair p. m. possibilities for these patients
00:16:59
um apart from surgery also we use a lot of radio therapy inky variation
00:17:04
and uh of course in the beginning everybody hope to these therapies would have the solution for these patients
00:17:09
so we would have less mobility but unfortunately that's not really turned out to be the case
00:17:16
uh we still use it a lot but we know now that uh that many of these bases also have big problems
00:17:21
um in the beginning of course you know the good side effects of where your therapy
00:17:26
but in the long run also a lot of things uh with this phone
00:17:30
yeah this phase yeah we're unique roses all these things happen and then i'm gonna
00:17:34
show you live examples depends very much you can imagine i mean really therapy
00:17:40
all days we just give them review therapy and feel from left and
00:17:44
right in front of nowadays they're really painted in because i'm marty
00:17:49
and the v. v. meant as all kind of modifications so you can really radiates
00:17:53
only part of the and you can try to to spare our part
00:17:58
and you can imagine so the part that you that you treat a
00:18:01
they give the most mobility so i'll swelling is affected um
00:18:06
i'm gonna show you a little bit although we are talking about swallowing it's nice to know how swallowing really works um
00:18:13
so so long is quite complicated because you have to chew and
00:18:16
it's a very complicated movements uh you learned have to lift
00:18:21
it will just has to go down your base of donors to touch the back one of the firings um
00:18:27
so and the larry's has to close so that the food goes into these of it's so
00:18:32
you can imagine for some patients in whom this mechanism does not really function anymore
00:18:37
and this is a billion after came radiation therapy for example in it doesn't really work very well
00:18:42
and then what we get is aspiration you sort of the foot went into the is offer
00:18:47
there's so it is in slow motion you see your it goes into the trade yeah
00:18:52
you should go into the is of this so and those patients sometimes have
00:18:56
really handicapped because they need a chip x. you look through the stomach
00:19:00
permanently to be fit to hold live uh it also has influence on articulation
00:19:07
and uh we know that the word of uh where all the the sounds uh are made in the mouse of course you can imagine that
00:19:14
the the area that has been a radiated most can really be
00:19:17
affected but we did some studies on that in our clinic
00:19:22
and uh it has been proven also to be the case that that the the high those
00:19:26
she uh uh on certain area schools uh articulatory problems in these patients
00:19:32
to give you an example is the patient uh uh is it that's
00:19:36
recording of a patient a before and after the key murray's therapy
00:19:42
well of him yeah i see it as
00:19:48
e. s. p. n. l. e. uh i uh
00:19:53
uh i like she will tell me
00:19:58
also this is a quite clear voice as you can imagine and now just after the game reduce their p. i. e.
00:20:08
sh t. i. e. t.
00:20:14
t. spanish train for instance so you see especially voices will change the station
00:20:20
i'm not a um so um although we tell the patients that we write is very user b.
00:20:27
we can spare their organs but it's not certainly not always the function that we can spare
00:20:33
um and we we recently looked one of our um our
00:20:37
residence at the voice quality ten years after chemotherapy
00:20:41
you can see the red bar is a deviant a parameter and the green is normal so very rough division of course
00:20:48
you can see all these parameters intelligibility degree buzz articulation nasality personally
00:20:54
almost all of them are influenced by by this therapy after ten years
00:21:01
uh well we did a lot of these kind of cynical we had the data of causes formula for more one of our recipes
00:21:07
she got all kind of outcomes like a roughness rather than as a steaming a strain page
00:21:13
at three or three months and and and and and we uh this this should
00:21:20
be post c. p. one post sorry thing and and this is three
00:21:24
months one one thousand and three one pose you can see we all these
00:21:27
parameters change so uh and mostly not in the direction of course yeah
00:21:35
well the the the testing these voice parameters eh is of course a lot of work for as a piece
00:21:40
and is very subjective like the previous speaker already said
00:21:43
so we are also looking for um objective measurements
00:21:47
so that can help the recipes and also the patience because it can get really score
00:21:53
so they get a score but a computer uh of the intelligibility on
00:21:56
this l. c. d. or what kind of aspect of the voice
00:22:00
and um we did this study and together we scanned an entrepreneur was uh
00:22:04
available who's also you in the audience and then so we looked
00:22:09
at for example this is there to the score by yeah
00:22:12
uh as a piece so that we're uh i think in this as though you were thirteen it'll
00:22:17
be that the the scoring of these patients are we compared it to the computer score
00:22:23
and uh you can see am it's certainly not the same always
00:22:27
but the trend is quite good especially in and this one running speech
00:22:32
um and um so well i think
00:22:36
a lot to improve still but of course we also don't know if the if the average of the use of these
00:22:41
physical standards that we use it as a gold standard to train the train the computer of course but um yeah
00:22:48
um apparently is difficult to get a to get a computer do exactly the same um but maybe the computer is
00:22:55
no for sure of course is that the computer is it can row b. doesn't always give the same score
00:23:00
and uh and uh and we can train them and we can twenty computer better
00:23:04
probably with more text but uh has said in the beginning of this
00:23:11
uh a little bit also butler reject me solar inject me is
00:23:15
uh when we are not the mouse uh it's treatable larynx
00:23:18
so it's more force probable and if we take out the
00:23:22
larynx um the patients lose the vocal cords so the
00:23:26
a much more difficulties to speech no three possibilities to do a rehabilitation is you have that
00:23:32
is a fragile speech so then the patient inject a little bit of enter into these office and group that up
00:23:38
and wait by burping you can make one two sometimes three syllables of sounds
00:23:43
um and so basic and do that very well but but it's quite difficult to learn
00:23:48
you have the electoral larynx probably many of you know that it's
00:23:52
in a electrical vibrating machine you keep to your neck
00:23:55
and it gives very monotonous aren't any or the trade you is a fragile speech and then a decision makes
00:24:01
a whole between the trade yeah and there is a for this and they put a a small fall
00:24:07
when they're so the foot without coming to do a trick here with the air can be pushed into these
00:24:12
office so then the page can make it sound with that so i'm gonna show you some examples
00:24:19
this is an uh electoral eric mostly now is it's in a electable vibrating machine
00:24:24
in the past they've also build into by several kind of or um
00:24:29
devices so so the it was less obvious that you had to can keep something in your mouth
00:24:34
there's not someone anymore nobody smoked by half or so and it doesn't really
00:24:39
help um and other devices have been designed also uh to do it
00:24:44
i will show you one uh design that's built into the dangers i think there is some future for this
00:24:50
uh if they improve would have eventually you it's not the problem with this is that it sounds very monotonous
00:24:57
so it's not nice to listen to just to let you hear in dutch sentence for this
00:25:09
so it's very monotonous not yeah uh so this is a air from
00:25:15
stacy machine they build it into the dangers with the remote control on the d. r. so now you you will hear that it
00:25:21
is still monotonous because it comes from the mouse any has a
00:25:25
microphone um it it sounds much better i think but
00:25:31
for the future this yeah growth
00:25:39
oh oh
00:25:44
i dunno novels are
00:25:55
oh so it's uh it's quite monotonous but it's uh of course very nice and to have to
00:26:05
but unfortunately it's made in the university of pennsylvania it's not commercially available um
00:26:13
so um what what changes if you have a trade you is a phase of was uh the thing we
00:26:19
mostly use nowadays uh in the world um a normal speech of course come from vibration of the vocal
00:26:25
cords and then in the in the speech can match it can now we ought to get articulate
00:26:30
what happened after nine inject me you you can agree to your mouth anymore bits with so much and if you make
00:26:36
its motions here you can direct the air through her hair again and indifference you can
00:26:42
get vibrations and vibrations can make the vibrations of your vocal chords now i'm gonna
00:26:49
like you see that's probably most of you a sinister was could be is supposed
00:26:52
to be is when you have a flickering light on the vocal chords
00:26:56
you can see the movements of the vocal cords and and that cause the
00:27:00
vibrations and cause the voice um and it's done because the um
00:27:05
the flickering light is just a little bit even from the frequency of the vibration that's why you can see it um
00:27:12
if you do the same with them the voice from the new very your new book is offering like me
00:27:17
you don't get these very nice vibrations but you can see it also but not was it's
00:27:21
tropical but was it very fast camera some gonna show you how is on looks
00:27:31
mm
00:27:41
mm
00:27:44
this awful inject me in the new ferries you can see this how how
00:27:53
wow you can see there are little more a regular uh
00:27:58
coming from all sides so the the sounds more rough not a not
00:28:02
a nice i'm not a nice uh uh harmonics out yep
00:28:09
uh i'm gonna let you hear some examples um sometimes you just take out the larynx but
00:28:14
sometimes also like the patient whether the longer section we have to reconstruct the variance
00:28:20
in this page it was reconstructed with a piece of skin from our our
00:28:24
and of course this all has influence on the voice because if you have a a piece of skin
00:28:30
in in your throat it doesn't vibrate as nicely for example as because that's off so um
00:28:39
yeah he's got the you are right human and because it should
00:28:44
be to you or that you can show you know
00:28:52
mm back to what you your voice you as you uh_huh six
00:28:58
do you um really don't william remote for me
00:29:05
h. ah yeah yeah yeah you in your
00:29:11
e. t. v. h. ah
00:29:17
you may be right you region yeah oh yeah oh well she said she
00:29:22
satisfied with the voice oh i'm in good although most of you
00:29:26
probably way she started voicing uh if you don't know than checking speech
00:29:30
or even my goodness this does not sound like a female
00:29:33
uh so especially for females of course it's it's quite a thing to happen that you get this very rough dark force
00:29:40
i'm at a telephone everybody will say mister so it's not nice uh they don't like it
00:29:46
nobody likes the voice often things like me think but it is of course of course
00:29:50
ah um this is basically the interface of voice but he's a very good is aphasia voice
00:29:57
it's in dutch but you will see a uh you can hear that it's quite fluent
00:30:02
um i'm gonna let you use some good and bad example this is a very good is if it's a voice example see it's a cat
00:30:14
it's it's
00:30:23
so we can make about five syllables on one of the little bit which is exceptionally good yeah
00:30:36
so i'm unfortunately not all people have that i'm gonna show you one that example
00:30:43
with the program you was we uh_huh alright first cousin on him and then
00:30:49
oh uh_huh nine i think i don't know
00:30:53
um then because of a dysfunctional levels yeah do you wouldn't unusable
00:31:00
you know and you're talking with i. u. vegetables yeah oh yes and i'm satisfied with your
00:31:08
voice ah ah ah ah ah ah ah ah because you like no more oh yeah
00:31:17
ah
00:31:19
ah
00:31:20
yeah and uh i pay she uh other people don't have to be used to listen to you or if you uh
00:31:28
because i mean uh i can really understand you know when you
00:31:31
talk but still um why don't you like your voice
00:31:37
ah
00:31:41
people cannot on it yeah yeah nothing spirit that uh well nobody would be very happy was such a force
00:31:49
uh on the other hand sometimes becomes very successful uh
00:31:54
one one example you know you you nobody where it will be oh
00:32:02
yeah and it's not the people who do you like what you have less
00:32:07
we know the voice means you kind of automatic speaking off so you don't have to put your finger and you can see
00:32:16
it's good it's useful yeah
00:32:22
yeah and i you can be used in la yeah
00:32:28
oh i know and you are you how long okay
00:32:33
um so recently uh we looked uh if in the literature there's really a good evidence
00:32:39
which voice is the best i mean in fact it's a little bit um disappointing because certainly not so there's
00:32:44
not so much literature really comparing all these uh electoral arises affect your voice tricky is value voice
00:32:51
and it would cost get a a a job to try to find the evidence that is out
00:32:55
there and these are some of the findings from uh from what she found so this is
00:32:59
on the um um a different frequency in a
00:33:06
sustained a or in a running speech
00:33:09
you can see this is healthy voiced recuse value was easy for your voice
00:33:13
you can see well there was a kind of tendency always that the trick
00:33:16
is a value voices better also in the studies these are filing plots
00:33:21
the same is with the foundation time much information time is affect your voice trek user for your voice a normal voice
00:33:27
how many to noise ratio of difficult to see any statistically significance
00:33:32
in this graph i think maybe also it kind of trends
00:33:37
and her shimmer and jitter also the trends
00:33:42
did you think that trick user face your voice is better than these effects of us something that we also would expect um
00:33:51
also uh we look at all kind of sit different surgical techniques uh even your call before my insecurity that so should
00:33:57
look at all kind of patients with two free flash step
00:34:00
back major flats standard injecting these guess should pull ups
00:34:04
a pack or or like major factor or she look at all kind of characteristics of these voices
00:34:10
and um so this frequency again but you can see
00:34:13
from these cross all the overlap um that um
00:34:19
that is difficult to conclude anything from this um and and maybe we can say
00:34:24
is that most of these bases can get a voice which is already nice
00:34:27
and in general the best voices either chewed free flaps or a standard lounge
00:34:32
activities and that also came out so the undertow this was um
00:34:39
um total percentage of speech so it is a common and this was
00:34:46
ah how much noise ratio and uh well all kind of voice parameters difficult to really
00:34:53
conclude anything from that i think a reasonably we looked in these patients if
00:34:59
uh and they after a long time still have the same voice quality or that
00:35:03
it improves or um or decreases so we do get patients in we had
00:35:09
recordings uh with the interval um uh of seven years or
00:35:13
more and we have thirteen patients in we had
00:35:17
really i'm a recordings and make recordings and um
00:35:23
uh it again in this we did it to experiments ten as a piece
00:35:28
try to look at the intelligibility of voice quality and we compare that to uh the uh it is just
00:35:33
to make and uh in a in belgium uh into a sister project that we get together with them
00:35:40
and um and it was the same sex so is it quite unique um uh material i think that so many
00:35:47
people have a recording seventy years apart from the same
00:35:50
patient group uh and um these were the findings
00:35:56
um so the um uh it if they are above the
00:36:00
lime then the new recording is better so the
00:36:04
the late after the surgery and i if it was below this line
00:36:08
the whole recording was better because it's difficult to we say
00:36:12
something from it's i don't think that there is a trend so um uh with that so they didn't find the trends
00:36:19
and it both ways intelligibility and voice quality because in some
00:36:22
patients uh it's really better and then or was was
00:36:26
and um the red one is there uh it s. range and do
00:36:32
a blue one is the one from the um as apiece
00:36:36
so you see sometimes they agree like here here but sometimes i also disagree
00:36:41
so again this machine learning a thing so this this and this automatic speaker recognition so certainly not
00:36:47
totally perfect so you would hope in this project also we get improvement um of that
00:36:54
uh so the conclusions already and told you so what will be the future in
00:37:00
lisa in design exactly page well nobody knows what the future is that
00:37:04
nice thing of course but one of the things that could happen we
00:37:07
hope it is the electronics will also start helping these patients
00:37:11
uh we now already know that's this is not for injecting patients for example
00:37:15
if we we have already computer application that you record your own voice
00:37:20
and then type in the text and the and the computer will will read
00:37:23
out the text with your voice so those things are report possible
00:37:27
and uh so that this could help these patients of course if they
00:37:31
keep their speech somebody in their voice could could talk about it
00:37:35
the things that one could think about is uh for example what kind of silence speech was monitoring of all the mouth movements
00:37:42
and that put into electronics into your own voice so you you would have to put in sensors in the tongue and lips
00:37:49
and uh well people are doing research in this field so this could be an application in the future
00:37:55
another thing could be just this vibrating then just like like i told you the like a show you
00:38:00
but then not with the remote control and the um with with with the sensor
00:38:04
built into to the away that record when they want to speak for example
00:38:09
and uh and then also not with with the the simple vibrating thing with with more a variation in speech
00:38:16
so that it is it sounds more natural the so those are things that we could think about for development in the future
00:38:23
um things we are working on at the moment cluster for example is on predicting speech
00:38:28
and also swallowing function and all these things do we we have a big program running in our institute
00:38:33
so that based on that human size location will treatment we're gonna get for your therapies surgery that
00:38:39
we can predict appreciable how well you will swallow how we sound how with voice will sound
00:38:45
now it's peaceful sounds so uh to be able to better counsel them and uh and also more directly
00:38:52
yeah person three weeks so that's was my last slide if anybody has any questions yeah well
00:39:07
oh yes
00:39:40
um well now i asked as i don't know so we have asked the patients so that's the first question
00:39:46
uh it depends very much on that disability if you have a little bit of uh articulatory problem of course because
00:39:51
you miss half your thumb like the the the gentleman i don't think you want to have any electronics
00:39:58
but um but if you hear some of these pages that hardly have a voice or our own was not
00:40:05
you can almost not uh i don't know i'm not intelligible of course they
00:40:09
are very motivated to have uh some electronic system hoping i'm sure yeah
00:40:22
yeah yeah yeah i'm a well even if you have a pool voice well the the very two voices they don't recognise
00:40:29
the voice anymore but but what is very striking even if you take out and there is because the accent
00:40:36
and uh there's still a unseen recognised errors as mister i also miss
00:40:41
huh because it sounds still a little bit the uh the same
00:40:45
yeah unless it's very pour of course yeah what is another uh yeah
00:40:50
it's quite expensive for uh well the the the rich rules they can after design inject image they can afford all the
00:40:56
four scorsese's because the page needs voice was is changed whatever's
00:41:00
every three months um but for example if you are
00:41:04
and and poorer countries even eh uh parts of your over there's no reimbursement
00:41:10
patient cannot afford it so they don't get nothing so the only have this is a fragile voice and
00:41:15
i showed one very good example but that's exceptional most patients is is the fact your voice
00:41:20
if you would offer them an electronic thing uh um that that that
00:41:25
functions well uh i think it would be very happy yeah
00:41:32
i i
00:41:39
i i
00:41:46
uh uh uh
00:41:58
yeah
00:42:01
i
00:42:11
oh
00:42:13
yep i don't need to buy hunky name oh so uh
00:42:23
i can't really projected over there but i'll show you here
00:42:26
so got my name is no trucks a home here
00:42:32
to you know i'm a a home
00:42:43
e.
00:42:46
based on a one on the back wall model e.
00:42:51
you don't ask ah ah ah fact you would be sort of that you come up
00:43:02
and ah that was like one i am not too far away from them like on what you do you will
00:43:13
still a little bit like what you were saying earlier on that uh people already have some tools that can
00:43:18
affect improvement in life or like the i will come back to your question
00:43:24
i think where he has to be contentious ease the adoption of it
00:43:30
people sometimes given the technology and if they had the ability to recall the what
00:43:34
and get a good quality voice like base the question isn't more about
00:43:39
whether i wanted all this voice or what i want to use another this the holes because people don't tend to
00:43:45
always want to use the white room on the all the disability that they've got so i think that you
00:43:52
to put a different slant to it but the technology is there
00:43:54
sometimes people don't always able to in the lane expected either
00:43:59
so i think yeah i it didn't want one to keep in mind that people technologies
00:44:04
that uh there are a little bit of a human element that we tend to
00:44:07
tend to overestimate absorption so we found that quite a
00:44:11
lot of our instances that uh we get about
00:44:16
between fifteen to twenty watches bit every week um so the
00:44:21
american university which builds these and they make this well
00:44:24
before this nine hundred dollars they used to make this fall prey this technology level for people to build watches
00:44:30
um but uh since we brought day technology into our at a local
00:44:35
is going so much that we can't afford to offer these
00:44:37
a full three anymore so they offer between fifteen and twenty which is a week but that's less
00:44:41
two hundred dollars it's quite achievable is quite accessible i agree with you that it's still available
00:44:47
mostly for western countries other than the some of the other uh um uh of developing countries well
00:44:53
but um at the what we found was that uh it was an element of uh
00:44:57
emotion involved besides the size to really put much to do with the solutions will
00:45:03
and we can commercially yeah it's available you can buy it

Share this talk: 


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.
302 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.
113 views
eTherapy
Elmar Nöth, Erlangen-Nürnberg
Sept. 25, 2018 · 11:45 a.m.
128 views
Assessment in speech disorders
Virginie Woisard, Toulouse
Sept. 25, 2018 · 2:45 p.m.
140 views

Recommended talks

Pitch Estimation Using Mutual Information
Yanbo Xu, University of Maryland College Park
Sept. 7, 2012 · 11:07 a.m.