Those Automated Voices Are People, Too

Bernie Wagenblast at the Newark Liberty International Airport AirTrain station, where his voice guides passengers every day. Photo by Liam La Guerre/LFRAG

Bernie Wagenblast at the Newark Liberty International Airport AirTrain station, where his voice guides passengers every day. Photo by Liam La Guerre/LFRAG

A version of this piece was published in Columbia News Service on March 4th, 2013

Karen Jacobsen probably says “recalculating” thousands of times a day. But she really said it only once. Jacobsen’s voice is used in millions of GPS units around the world, as the “Australian Karen” option. But she wants to take people beyond that missed exit on the highway, or their final destination.  She created a brand called GPS Girl that works to inspire people.

Whether it’s on GPS, phones or trains, many voice actors want to connect to their listeners beyond their recorded phrases and input their personalities in the works.

“Not only do I give them directions in cars, I give them directions in life,” Jacobsen said. She wants people to know that  “they can be in the driver’s seat of their own lives and that they can recalculate in life at anytime, like they can in their car.”

The voices in our devices can be human, computer-generated or humans whose words are spliced together in ways that can sound computer-generated. But even as computer-generated voices become more common in devices, many companies said they preferred recorded human actors because, well, they sound more human.

“The thing that a computer will never be able to do is reproduce is a personality,” said Andrew Breen, a director at Nuance Communications, a speech technology company. “If we copied William Shatner’s voice, it would sound like him, but it wouldn’t be a performance that William Shatner gave. It would sound like a bad actor trying to be William Shatner.”

Bernie Wagenblast at the Newark Liberty International Airport AirTrain station, where his voice guides passengers every day. (Photo by Liam La Guerre/CNS)

And behind those personalities are real people. Bernie Wagenblast is one of the voices used in the New York City subway system.  His recorded voice booms in some stations “There is a downtown 1 train approaching the station — please stand away from the platform edge.”

He’s not a newcomer, nor is his voice unfamiliar to New York commuters. A New Jersey native, Wagenblast began his career in 1979 as one of the original traffic reporters for Shadow Traffic, which gave traffic reports for local stations that didn’t have their own at the time, such as WABC and WINS.

After Shadow Traffic he worked in communications for New York City’s Department of Transportation and then as an operations manager for the Port Authority of New York and New Jersey.

In the early 2000s he was asked to be the voice of the AirTrains at Newark Liberty and then John F. Kennedy International Airports. He also recorded a commercial for the New Jersey Transit.

Wagenblast ended his hiatus from traffic reporting about five years ago and now can be heard on WINS on Saturdays.

About the same time, the Metropolitan Transportation Authority asked him to do the voice work for the subway. “Being a transportation person as well as a voice person, I said ‘sure this sounds like a lot of fun,’” he recalled.

He was emailed a spreadsheet of about 1,000 words and phrases and went to a recording studio to repeat them. Words such as “uptown,” “downtown” and “southbound,” and he recorded numbers from “first” and “second” to “242 Street.” Everything had to have the same tone or he would have to repeat it.

“By the time you get up to about 154 Street your eyes start to get crossed and you become a little punch drunk,” he said.

When a train is arriving or approaching the station, a computer takes his individual recorded phrases and puts them together to make complete sentences.

“Because I’m from the New York area most of the words I already knew how to pronounce,” Wagenblast said. “Take the Van Wyck Expressway. Some people prefer ‘Van Wyck,’” rhyming with “like,” “but other people prefer ‘Van Wyck,’” rhyming with “stick.” But he noted that years of voice training resulted in the clear, booming mid-American accent he has.

Susan Berkley, the voice of call centers for Citibank and AT&T, said her method to connecting with listeners is to imagine them before her.

“I am always visualizing and feeling the presence of another person who isn’t there whenever I do voice-overs,” Berkley said. “It’s very deliberate and part of voice acting technique.”

Berkley, who has been a voice actress for more than 25 years and is the president of the Great Voice Company, added, “Body language, not just facial expression, makes a huge difference.”

For a large project speech project that IBM was doing based on her voice, Berkley was recorded reading various books, random emails and phrases like “for help at any time press zero.” The entire project took 50 hours to complete, which she did in four-hour sections a few days a week for months.

But even though injecting personality in the automated voices helps connect with others, sometimes a live human being is necessary, as Berkley knows from both sides.

Berkley was the victim of identity theft in July 2000. When she called AT&T’s billing department, she didn’t get a human operator at first.

“When I called to tell them it was identify theft I got my own voice,” Berkley said. “When I finally got a human being I was trying to explain to them that I was one of their voices and I thought ‘No don’t go down that road, they’re not going to take me seriously.’ Then I called the credit reporting agency and … it was my voice.”


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: