The science of speaking

"SO speak to me in Selby," says Professor David Howard. "What is your real accent like?"

I've been unwise enough to admit that the voice' I use for work is different from my native Selby accent - and it's one that I'm not particularly comfortable with.

Big mistake. Because this man just happens to be an expert on the human voice. He helped Rory Bremner improve his Tony Blair impression, analysed the famous coughs from ITV's Who Wants To Be A Millionaire, and can often sense your state of mind just by listening to you speak.

The professor of music and technology at the University of York also hosts a BBC4 programme called The Voice, which gets a second airing tonight on BBC4 at 8pm.

And now he wants to turn his analytical powers on me.

At work, I prefer to "put on" an accent, to sound more professional. My true colours only really come through if I speak to one of my sisters.

David suspects that I disguise my accent by pronouncing more consonants, and to prove it, we record my voice on to a spectrogram.

He instructs me to say "my name is Charlotte and I work for The Press newspaper" in both my work and Selby voices, while strong black lines strike up and down the screen to represent my speech.

The results are fascinating.

My "work" voice is almost half the speed of my natural accent, observes David, and sounds lower.

That could be because I have to think about it as I go along, but it also fits into a theory devised by author Anne Karpf, who believes that women's voices have lowered over the years, as their careers take off.

"She says it's to do with the increasing number of women that go into the boardroom," said David.

"Margaret Thatcher is a prime example of that. If you hear her when she was an MP then at the beginning of her premiership and at the end, there are steep changes in her voice. It gets lower and lower and says I am calm, I am collected and I am in control'."

So I'm unconsciously doing a Maggie at work, then?

The way I pitch my voice at work could well be because I want to sound more assured and in control in my professional environment - both for myself and those around me, says David.

Consonants are another matter, and there are many reasons why you might enunciate them well, he explains.

"For example, if you and I were talking intimately, our consonants would disappear because we wouldn't want to share our conversation with anybody near us.

"Our speech would go down in level and everything would become less articulate. But if I was way over there, your consonants would be fantastic. If there was an emergency and you wanted to tell me something I would get the message loud and clear and the whole room would know.

"You wouldn't have to think Selby or York, you would just do it. That's a feature of our speech - it changes all the time."

He should know.

This is, after all, the man who, in addition to analysing those "Millionaire" coughs, created a Premiership league table based on the tunefulness of fans' singing.

When he brought the table out in 2002, Southampton, then struggling in the Premiership's lower reaches, came top.

His research also proved that there really is no difference in the quality of sound produced by an all-girls' and all-boys' choir - debunking centuries of choral tradition.

His job as a professor of music and technology is to analyse singing and teach music technology.

But he has also been "bought out" of teaching by the Engineering And Physical Sciences Research Council so he can devote time to getting more young people interested in science.

He also leads the media engineering research group, which focuses on television image processing, making computer and video games more realistic, specialised audio production equipment and voice production.

That was how he became involved with the BBC in making tonight's programme.

In it David strives to show us not only how our voice works, but what it says about us and our background, gender, age, education and emotional state.

It was a lot of fun and extremely interesting, says David, but it had a practical purpose too.

"One reason for doing that is so we can make computers speak," he said. "The main issue today is how can we make it sound natural? Like a human being rather than a computer. For instance, when you ring up directory inquiries, they play the number you have asked for back with a synthesised voice.

"Synthesisers are part of every day life but we know they're computers - we don't mistake them for a human.

"The programme starts with the premise that if you're trying to create electronically natural-sounding speech, we need to find out more about how the human voice works.

"At the moment, the models used in synthesis are doing everything they can. We have reached the end of what we can do with them and the new models need new thinking. The programme says lets go out and find out what humans can do with their voices."

The programme makers let David choose his own itinerary, and he invited comedians Rory Bremner and Jeremy Hardy to help him - even taking up the challenge of helping Rory improve his impression of Tony Blair.

"That was a tall order," he said. "For starters, although they said I had three hours, about two hours and 40 minutes of that was spent messing about with make-up, lighting and cameras."

Later, they recorded Rory's speech on a spectrogram to get a visual representation of it and compared it with the former prime minister's. They discovered Rory's consonants were not as pronounced, which gave him scope to improve his impression.

"But afterwards, he came out with what I thought was a revelation," enthuses David.

"He said that he has a different set of teeth in for his Tony Blair impression, behind his own teeth so you can't see them, but if you think about where you tongue is when you make the tuh' sound, Rory's tongue will obviously be somewhere else if he has an extra pair of teeth."

So back to my voice.

If you listen to it, it does - I hope - give away that I am a woman in my twenties.

If you listened to me again in my seventies without being able to see me, it could however be much harder to tell whether I was a man or a woman, David says. That is because a boy's voice deepens by about an octave when his larynx starts to grow in his teens.

A girl's larynx doesn't grow anywhere near as much - so her voice will only lower a little. A young man's voice, therefore is generally much deeper than a young woman's.

Interestingly, however, things change again from your mid sixties, to the extent that when men and women reach their nineties, their voices will end up at around the same pitch, David says.

From listening to my voice, he says I am emotionally content. I tell him I would not bring personal problems to work anyway, but David is not convinced. Our voices give much away, he says, and you could tell within ten seconds of a phone conversation whether someone was upset or not.

"If you're upset about something that has nothing to do with this situation it would be very difficult to hide that in your voice," he said. "When you're upset about something, your voice will almost certainly have a catch.

"When you're upset about something it's very difficult to speak slowly and not have a slightly higher pitch.

"Your voice is a fabulous way of communicating your feelings and it carries your emotion all the time. It doesn't matter what state you are in, it gives you away."

When it comes to your voice, it seems, you can take the girl out of Selby, but you can't take Selby out of the girl.

Voice facts

* Female opera singers are not encouraged to sing heavy opera (such as Verdi), until their 30s, because the tissues of the larynx won’t have settled down enough.

* Cockatoos can mimic the human voice. Some clever cockatoos can mimic up to 800 words.

* When we are born we all have the same voice pitch – you can’t tell a baby’s gender by listening to it cry

* When female singers are in their teens, they might lose the top few notes of their range for about six months, but they will get them back when the larynx has finished growing

The Voice

Professor Howard's hour-long documentary on the human voice will screen again tonight at 7.30pm on BBC4.

The programme, which features the choristers of York Minster, comedian Jeremy Hardy, tenor Alfie Boe and jazz singer Jacqui Dankworth, starts by asking why electronically synthesized speech is easy to understand, but not very natural-sounding.

Prof Howard, an expert in electronic voice synthesis, then sets out to push the limits of voice synthesis using the latest technology.

Along the way, he meets Pavarobotti, a singing robot that duets with its inventor, and gives voice production tips to Hardy, a regular on Radio 4 comedy programmes, known for his tuneless singing, and tenor Alfie Boe.

The programme also analyses major political speeches of the 20th century, investigating how speakers engage with different audiences, and considers whether a woman's voice is likely to drop in pitch as she becomes more senior in the business world.

The thorny issue of overseas call centres, and the changeover of Britain' speaking clock, are also under the microscope, while the programme starts with the choristers of York Minster and ends with a special octet drawn from the Beningbrough and Ebor Singers as a backing group for opera, jazz and gospel singers, to highlight an extreme of human vocal achievement.