If I tell you:
“The voice is unique to humans,”
you will all tell me: “Yeah, obviously.”
Well it isn’t,
the voice is not unique to humans, the voice is everywhere.
The word is unique to humans.
But the voice is everywhere.
All the animals have a voice: mammal, terrestrial or marine animals,
all the birds, obviously,
even some amphibians have a voice.
We also understand perfectly
the intonations of our four-legged companions.
Indeed we say that they only lack the words, not the voice!
They do have the voice.
The voice is not unique to humans.
The voice is everywhere.
So much so that
it has now entered the technology of our everyday life.
And it seemed so natural that we hardly realized it.
Today, the call platforms
make us interact vocally with answering machines,
the voice of our GPS tells us the way to go
and we can even have a conversation
with our mobile phone, can’t we?
Synthetic voice: I do not doubt it, Johanna.
JR: How are you?
Voice: Hunky Dory.
JR: Great.
What you see here is not a Lady Gaga concert.
Alys is her producers’ little darling.
She is never sick, never sings out of key,
and she does not ravage her hotel room.
Because Alys is a hologram.
She is computer-programmed
and her voice is as virtual as her presence on stage.
(Music)
It’s crazy, right?
All these voices are synthetic voices.
They are perfectly intelligible,
extremely powerful and better still, they are infallible.
Yet, they remain perfectly recognizable.
Voice: Hum, hum, sorry, but I do not agree.
JR: Yes, sorry, I forgot to introduce you to someone.
The voice you’ve just heard is a synthetic voice
that I programmed specifically for this TEDx.
I called it HAL as a tribute to Stanley Kubrick’s computer.
Voice: I love Stanley Kubrick.
JR: We’re very happy to hear that, Hal,
but like it or not, your role today
is to demonstrate the gap that exists between us.
And more specifically, between your synthetic voice and my human voice.
Voice: I do not know what you mean.
JR: Well, I’m getting there.
What separates us is not what we say,
that’s how we say it.
Let’s take an example:
could you, HAL, read the sentence on the screen, please?
Voice: Gladly; I did not say you stole the money.
JR: Great. Now repeat it insisting on the fact
that it was not you who said it.
Voice: Easy. I did not say you stole the money.
JR: Well, that’s what I meant when I was talking about a gap between us.
The synthetic voice has no intention.
It is neutral, in all circumstances.
Moreover, if we look at the curves representing the melody of its voice,
we realize that between the two productions,
it did exactly the same thing.
While I could say:
I did not say that you stole the money.
I did not SAY that you stole the money.
I did not say that YOU stole the money.
And just with the sound of my voice,
not only will I give a different meaning
to a sentence that contains identical words,
but also, I will convey all my intentions.
So when you compare my curves with HAL’s,
you realize that whenever I want to insist on a word,
my voice becomes louder, more high-pitched,
and it is immediately followed by a pause.
But HAL is not the only one
for whom the transmission of intentions is difficult.
Even the most powerful computers have difficulties
and are limited by their prosody.
John Oliver: Be honest, is that really the highest honor you’ve ever received?
Stephen Hawking: Yes.
JO: Well, discerning your tone of voice is a bit difficult.
By saying that, are you being sarcastic?
SH: Yes.
Voice: Okay, but can’t it be programmed?
JR: Yes, technically, it’s feasible.
On the other hand, what I will never be able to give it is a personality.
Because beyond the sounds that come out of our mouth,
what makes our voice human
is it’s going to convey all the most intimate aspects
of our personality.
Our voice says everything about us.
At the moment we open our mouth,
our interlocutor immediately decodes information about our identity,
and in particular, our gender and our age.
Why?
Well, all or almost all at the beginning depends on the size of the larynx.
By the way, when Titze creates a mathematical model
of a very small larynx, with 2 mm vocal cords,
and that he plays it on his computer, here is what it is like.
(Baby crying)
Then, he takes exactly the same model without changing anything at all,
except the size, to get 5 cm vocal cords,
and that’s what you get.
(Lion roar)
Titze deduces that the only difference between the two
is the size of the vocal cords.
In humans, it works in pretty much the same way,
that’s what will allow us to make a difference
between a man’s voice and a woman’s voice,
an adult voice and a child’s voice.
Finally, most of the time, since to any rule,
there are always exceptions.
What we see on this curve,
is that there is a common area for men’s voices and women’s voices,
between 145 and 165 hertz.
Within this area,
there are high-pitched men’s voices and low-pitched women’s voices.
In my practice as a vocologist, I treat broken voices
or I optimize the voice of those with specific needs:
lawyers, lecturers, business managers.
When I work, for example, on feminizing a transsexual’s voice,
that is to say, a biological man who lives his life as a woman,
this is the area I’m trying to reach.
By the way, I will perhaps surprise you if I tell you
that we are all capable, at the mere listening of a voice,
to estimate the age of its speaker, more or less five years.
Which is incredible as a level of precision.
From birth to adulthood, it’s a matter of growth,
so always a matter of size of the larynx.
Then it will be a matter of aging.
It’s not magic, vocal cords age like the rest of the body.
Muscles lose tonicity, volume, and the voice gets breathy.
Hal’s voice has neither age nor sex.
The only thing we can say is that it’s an adult’s voice.
Neither a child’s nor an elderly person’s.
Hal: I do not know what you mean.
JR: By the way, when measuring the pitch of its voice,
we realize that it is 160 hertz.
Right in the ambiguity zone.
So, no sex, no age,
HAL’s voice has no personality and does not convey any emotion,
because it doesn’t feel any, precisely.
When we think of emotions,
we always think about their psychic dimension,
but emotions have a physical, biochemical starting point,
quite concrete inside our body.
For example, when you’re scared, there’s an adrenaline rush
that makes your heart beat faster, your breathing accelerates
and all the muscles in your body get tense
to be ready to face the situation.
Well, as our larynx uses both
the air of our lungs and the muscles of our vocal cords,
it can be heard in the voice.
Because it also conveys emotions,
our voice is a real sound face.
That of our loved ones conveys the whole relationship we have with them,
all the heritage of our common history.
That voice,
Voice: He lost them all in the same way.
JR: It doesn’t remind you of anything, of course.
But this voice, it moves me because it is my grandmother’s voice.
When I hear it, it all comes back to me:
my holidays at her house, our complicity, her smile
and, of course, her face.
All this, technology is not able to reproduce it.
In your opinion, why, otherwise, almost 50 years apart,
these two directors would have chosen real actors’ voices
to “embody” the voice of their computers?
For Kubrick in ’68, one can easily imagine that he did not have the technology.
But for Spike Jones in 2014,
the choice of Scarlett Johansson is perfectly clear.
Because when one wonders about the reality of things
and their representation,
then you have to consider that…
Voice: This is not a voice.
JR: It’s a word-speaking machine.
Voice: But then you mean
that I will never be able to be a real voice?
JR: Research is making progress,
technology too, at a hallucinating speed.
The young woman you see here, and her incredible realism,
able to have facial expressions of emotions, is actually a robot.
Already, she is programmed to respond to a real actress.
Tomorrow, with the progress of artificial intelligence,
she may be capable of free will,
therefore of subtle and autonomous interaction with human beings.
Tomorrow she will be your perfect secretary.
Tomorrow, she will be your perfect babysitter.
Tomorrow, she may be the woman of your life.
Voice: Or a TEDx speaker.
JR: Or a TEDx speaker.
(Applause)
”
Be First to Comment