No more Dalek voices! Creating authentic voices globally, insights with Dr Matthew Aylett,

No more Dalek voices! Creating authentic voices globally, insights with Dr Matthew Aylett,

Interesting interview with Dr Matthew Aylett, Chief Scientific Officer at CereProc

What is your background briefly?

I’m currently the Chief Scientific Officer at CereProc and I’ve worked in the synthetic speech industry for around twenty years, co-founding CereProc in 2006. I studied Computing and AI at the University of Sussex before moving to Edinburgh for my MSc and PhD, the latter of which explored how and why we say the same words differently in different contexts.

Does it seem like a logical background to what you do now?

Absolutely – but interestingly, it wasn’t until my sister took on a project of her own that I realised my real passion and mission. She was carrying out a project around bullying and asked me to create five expressive voices to help with the research. The technology at the time wasn’t set up to make synthetic voices with emotion, so it was an interesting challenge! After a short spell at the International Computer Science Institute in California, I decided to team up with Chris Pidcock and Paul Welham to turn this challenge into a commercial company, creating CereProc.

1 min pitch for what you are doing now?

We’re creating the best, most characterful and expressive synthetic voices in the world. Accented voices, young voices, old voices, regional voices, emotional voices – if you lose your voice because of an illness like Motor Neurone Disease, you don’t want to sound like a Dalek for the rest of your life. You want to sound like you.

Why did CereProc develop an Irish TTS system and why do you offer an Irish accented system?

We’ve been involved with a number of European Commission projects over the years, and recently started working with the Grassroots Wavelength project, which is aimed at reinvigorating community ties through local radio. One of the project sites is Bere Island, in South West Cork. They are exploring how to run a radio station for a few hundred listeners and how to support presenters who have limited time and resources. We developed the Irish voice to fill in the gaps with pre-written content that could be spoken with an Irish accent; it would have completely alienated listeners to tune in and hear a synthetic voice that sounded like the Queen!

Can TTS ever show emotions? Should it?

Absolutely – we’ve been able to do this for some years. It’s particularly important when you’re helping people who have lost their voices through illness. For example, the roboticist Peter Scott-Morgan and the US news anchor Jamie Dupree both lost their ability to speak through different illnesses and we’ve been helping them regain their voices through text-to-speech. It’s vital to their identities that they can express emotions with their voices; without it, it’s much harder to express yourself clearly – or just crack a joke!

How will this technology change in the future?

Early TTS systems used a system called Unit Selection, which recorded a large amount of vocal content – around thirty hours – before effectively chopping it up and smoothing it out. Since then we’ve accelerated this process using neural networks and the newest systems, in the style of WaveNet products, produce extremely fluid synthetic speech extremely quickly. Future technology will only improve on this, helping more disadvantaged people to speak more naturally and fluently.

It’s important to note that these developments don’t just apply to radio stations and people who have lost their voices – we’ve also seen very promising projects in the gaming and entertainment industry (one of our voices is used in an escape room, for example), virtual assistants and even historical re-enactment – we re-created JFK’s voice to deliver the speech he never could.

In recent years, it seems that the internet and social networks have been flattening and homogenising the globe, rather than helping us communicate and preserve the individual differences that make up our unique cultures and personalities. We hope that CereProc’s future technologies can help to correct this, creating regional and expressive accents for everyone across the world. It’s the beginning of a very exciting time for synthetic voice.

How can people find out more about you & your work?

You can follow CereProc on Twitter @CereProc, visit our website at CereProc.com or find us on LinkedIn or YouTube.

More information about Irish Tech News and the Business Showcase

FYI the ROI for you is => Irish Tech News now get over 1.5 million monthly views, and up to 900k monthly unique visitors, from over 160 countries. We have over 860,000 relevant followers on Twitter on our various accounts & were recently described as Ireland’s leading online tech news site and Ireland’s answer to TechCrunch, so we can offer you a good audience!

Since introducing desktop notifications a short time ago, which notify readers directly in their browser of new articles being published, over 16000 people have now signed up to receive them ensuring they are instantly kept up to date on all our latest content. Desktop notifications offer a unique method of serving content directly to verified readers and bypass the issue of content getting lost in people’s crowded news feeds.

Drop us a line if you want to be featured, guest post, suggest a possible interview, or just let us know what you would like to see more of in our future articles. We’re always open to new and interesting suggestions for informative and different articles. Contact us, by email, twitter or whatever social media works for you.

Home

If you would like to have your company featured in the Irish Tech News Business Showcase, get in contact with us at [email protected] or on Twitter: @SimonCocking