AI gives silenced radio journalist his voice backNews 

AI gives silenced radio journalist his voice back


AI gives silenced radio journalist his voice back

US radio journalist Jamie Dupree Symbol copyright Jamie Dupree Image caption Radio journalist Jamie Dupree can be capable of broadcast again way to synthetic intelligence

A US radio journalist who misplaced his voice years in the past will quickly return to the air, way to synthetic intelligence.

Jamie Dupree, 54, a political radio journalist with Cox Media Crew, is unable to talk because of an extraordinary neurological condition.

a new voice was once created for him by Scottish era company CereProc.


CereProc skilled a neural community to predict how Mr Dupree may communicate, the usage of samples from his vintage voice recordings.

“This has saved my process and saved my circle of relatives from a terrible financial unknown,” Mr Dupree told the BBC. “there may be not a lot of a marketplace for radio journalists who can not communicate.”

Image copyright CereProc Symbol caption Users report themselves reading out a series of sentences, which is able to be then became a voice

The neural networks, which include among six to ten layers each and every, work by means of reducing audio recordings of phrases right down to phonetics.

The synthetic intelligence machine slices each and every word learn out via an individual into ONE HUNDRED tiny pieces, and does this with quite a lot of common phrases until it understands how elementary phonetics paintings in that particular person’s voice and has an ordered sequence for the entire items in each and every phrase.

Then, the neural community can create its own sounds and predict what the individual might sound like in the event that they have been to say a chain of words in conversation.

Many computer scientists across the international are trying to replicate the human mind by means of training neural networks to perform image reputation, however CereProc says that it is much easier to use artificial intelligence to sound.

“AI techniques work quite well on small limited issues, and learning to fashion speech is one thing deep neural nets can do very well,” Chris Pidcock, CereProc’s chief technical officer and co-founder, informed the BBC.

“it’s a much more solvable drawback than gadget intelligence.”

Silenced through illness

Mr Dupree has been overlaying political information from Congress in Washington DC for the previous 35 years. And as a journalist producing content material for 6 radio stations, his voice is essential to his paintings.

He started shedding his voice in 2016, but there has been nothing unsuitable along with his vocal cords, throat or larynx.

After baffling doctors from several massive US university hospitals, eventually Mr Dupree was once diagnosed with tongue protrusion dystonia – a unprecedented neurological situation where the tongue pushes forward out of his mouth and his throat tightens each time he wants to speak, making it not possible for him to mention greater than or 3 words at a time.

Rather than give up his work, Mr Dupree endured to do interviews with policymakers in Congress using an eWriter tablet to scribble questions all through one-to-one interviews, or via recording the solutions given to teams of newshounds in the Senate building’s hallways among hearings.

Image copyright Jamie Dupree Image caption Jamie Dupree at paintings in a radio studio

Even If he used to be still writing and generating stories, he had necessarily long gone off the air completely, because he could not present the stories he had written.

Then, in December, a member of the u.s. Congress spoke out on his behalf at the flooring of the home of Representatives.

The ensuing media consideration spurred his enterprise to take a look at to seek out some way for Mr Dupree to return to the air, because it had virtually 30 years’ value of his radio broadcasts on record.

a brand new voice

way to the computer-generated voice produced via CereProc, from Monday, 25 June, onwards Mr Dupree will once again be heard by means of WSB Atlanta listeners, to boot as audiences of Cox Media-owned stations in Orlando, Jacksonville, Dayton and Tulsa.

With His new voice, Mr Dupree can now write a script after which use a free text-to-speech software referred to as Balabolka on his laptop to show it into an audio recording.

If a phrase or turn of phrase doesn’t sound rather proper in the recording, he can sluggish certain consonants or vowels down, or change a word to at least one that does paintings, or amendment the pitch, and he will have an entire radio tale able to move reside in only seven mins.

“it is me, there is indisputably approximately that,” said Mr Dupree.

“Yes, it is reasonably robotic, but no-one used to be promising me that it was once going to be very best.”

In individual, whilst speaking to circle of relatives and co-workers, Mr Dupree nonetheless has to depend on the eWriter tablet, or saying a pair of phrases slowly, however the new voice has made a large difference to his existence.

“that is superior,” he mentioned. “Writing for my blog, sending out tweets and doing Facebook is superb – but there may be not anything like cranking out a 20-second tale jammed with a pair of sound bites to make the highest of the hour newscast.”


Related posts

Leave a Comment