A new software is helping children with speech disorders

The word “euphony” means pleasant sounding to the ear. It’s also the name of a startup that’s using innovative software to help children on the autism spectrum find their voices.


'Euphony' -- it's a word that means pleasant-sounding to the ear.

It's also the name of a startup that's using innovative software to help children on the autism spectrum find their voices.

Here's a look.

Euphony's got an interesting story behind it.

We were working on a project for a government contractor in the advanced-research division, and the government sequester and shutdown at the time put in jeopardy a lot of the follow-on funding that we were really relying on.

Consequently, all of our projects got shut down.

So, I took that as an opportunity, instead of just a chance to lament, and asked my company if I could take out the technology that we'd actually developed that would have ended up rotting on a shelf.

So, we formed a small business, and here we are.

Really exposed me to a different world, where speech is important, and that's where people either don't have the voice they want or don't have a voice at all.

So, people within the speech-communication-disorder community, autism-spectrum-disorder community decided that that would be a really neat place to bring our technology advances and make use of them.

What we have managed to focus on specifically right now, looking at how to model emotion and context change, has allowed us to do the equivalent of putting a smile onto a face that doesn't smile -- doing that with a voice.

[ Monotone ] He turned sharply and faced Gregson across the table.

[ Inflected ] He turned sharply and faced Gregson across the table.

Right now what we've done is we're partnered with a speech-language-pathology company to replace the speech capability in one of their apps.

And that's called 'InnerVoice.'

InnerVoice is a mobile app.

It's available now on the Apple iTunes store.

It's only on IOS currently.

And it is a communication app.

So, if you have challenges with communicating, talking, with language, social language, this app not only can talk for you, but will teach you how to talk.

I want hot chocolate.

And Fuzz is the expert there.

He's working on all these wonderful text-to-speech things where you can put emotions into text-to-speech, and now we can teach our kids how to say things maybe happy.

Today it's fairly robotic.

[ Monotone ] I want an apple pie.

And it doesn't connect the emotional content with the user.

It doesn't say, 'This is what you should sound like when you're happy or sad.'

That's a piece that's missing.

So, what we're focused on right now in this project is adding three significant emotions and tying those to improvements to the face.

State of the art.

[ Monotone ] I like having a voice to speak with.

Here's Euphony's version of that.

[ Inflected ] I like having a voice to speak with.

It's a running spectrogram.

So, it's recording everything I'm saying and displaying it in a spectrogram that measures frequencies, frequency ranges, and power -- those kinds of things.

And it's live.

But what it really demonstrates is how much is involved in analyzing a speech signal.

You can see all the noise up top.

If I make 'S' sounds -- 'ssss'.... That's what I'm analyzing under the hood.

When I make voice sounds, like vowels in particular -- 'eeee' and 'aaaa' -- you can see what changes here based on the shape of the mouth.

The biggest challenge is how do we build the ability to change emotional context without completely overhauling and rebuilding a synthesis engine from scratch?

As a startup business, I don't have the resources to do that.

So, what we have successfully demonstrated is that we can build voices that very effectively drop into existing technology environments with zero or very few changes in some cases and allows a new context to be introduced.

It's completely out there.

The other companies are great.

They've been creating the grid styles since the '60s.

And, truthfully, when mobile technology came out, they took that same style and put it onto the mobile technology.

Well, we know mobile technology is capable of so much more than that.

It's wonderful.

I want a cookie.

The reason Euphony decided to really focus on this is because numbers in different areas are changing.

For instance, the 1 in 68 people being diagnosed with autism.

We know that's changing.

And that's just a number we're familiar with in the Western world, English-speaking world, predominantly.

This is a worldwide problem where 1 in 68 people could potentially not have the same opportunity to communicate in life verbally.

Should have that opportunity if they want it.

Not everybody necessarily wants to express themselves with speech, and that's fine.

But for those who want to, they should be part of our culture.

I want a cookie.