海角大神

Samantha Laine Perfa:聽Welcome to 鈥淩ethinking the News.鈥� We鈥檝e got the final episode of our series, 鈥淪ay That Again?鈥� hosted by the Monitor鈥檚 Jessica Mendoza and Jingnan Peng. This whole season they鈥檝e focused on how we speak and what that says about who we are. And it鈥檚 true, having a 鈥渧oice鈥� is critical. But identity and personal agency are about more than just our natural ability to speak. Take a listen.

Jingnan Peng:聽This is Larry. He lives in Canberra, Australia. And this recording was made by his wife, Alicia, just moments before an operation that doctors said would very likely impair his ability to use his voice.

Jessica Mendoza:聽Three weeks earlier, Larry was diagnosed with cancer in his mouth. The sudden timing meant that he and Alicia had to rush to prepare for this huge change in their lives. And they didn't really have time to mourn for his voice.聽

Peng:聽Larry and Alicia have been married for about 30 years. Just before Larry went in for his surgery, Alicia managed to record Larry telling her favorite jokes, and singing one of her favorite songs, 鈥淟a mer.鈥�

Peng:聽You鈥檙e listening to 鈥淪ay That Again?鈥� 鈥� a podcast about how we sound, how we listen, and why that matters. From 海角大神, I鈥檓 Jingnan Peng.聽

Mendoza:聽And I鈥檓 Jessica Mendoza.聽

[THEME MUSIC]

Peng:聽So, this is our last episode of the season.聽

惭别苍诲辞锄补:听I can鈥檛 believe it!

笔别苍驳:听Yeah. And you know, all throughout this podcast we鈥檝e been talking about the idea that our accents, languages, and voices 鈥� they shape and reflect who we are. But what about people who have trouble using their voices?聽

Mendoza:聽Right. Today, in the U.S. alone, some聽聽people belong in that category. In previous episodes, we talked about how so much of being human is wrapped up in communicating and expressing ourselves. And so this can be a big challenge for folks with speech disabilities.

Peng:聽Thankfully, advances in technology are allowing more people to express themselves. With voices that are more and more human-like and personalized.聽

Mendoza:聽Today we share two stories of people who are turning to this technology for help. For Larry, it鈥檚 about preserving his own voice. And for Sean 鈥� you鈥檒l meet him later 鈥� it鈥檚 about creating one he has never had.聽

Peng:聽In their own way, each is navigating the question: how much can technology help us construct our voice, and shape who we are?聽

Mendoza:聽This is Episode 6: To Build A Voice.聽聽

[MUSIC]

Peng:聽I spoke with Alicia over Zoom about two weeks after the surgery. Larry was still recovering, and only made a brief appearance halfway through the call.

Mendoza:聽And you actually thought Alicia was Australian at first, right?

笔别苍驳:听Yeah, from her accent. But I found out she鈥檚 from the U.S.

Mendoza:聽Jing, you know you鈥檙e not supposed to make assumptions based on the way people speak.

Peng:聽I know. You're right! After reporting a whole season about this, here I am still.聽

惭别苍诲辞锄补:听Larry and Alicia met in the early 1990s, while they were both stationed at an army base in the U.S. Larry is Australian, and the couple moved to Australia soon afterwards. Today, they both work in risk management.

Peng:聽They asked us not to use their last names, because they're going through a private time and don't want to be contacted about their story.

Mendoza:聽Larry has been diagnosed with cancer once before, about 10 years ago. And he survived it.聽聽

Peng:聽But in early February, about a month before our Zoom call, Larry got the news that the cancer had come back. This time, the doctors said:聽

笔别苍驳:听This is Rupal Patel.聽

Peng:聽VocaliD builds synthetic voices for businesses that use them for things like dubbing video content, or answering customers鈥� calls. Voices that sound like this:聽

Mendoza:聽But the company also creates customized voices for people with speech disabilities.聽

Peng:聽So a lot of folks with speech disabilities use text-to-speech devices that turn phrases you input into spoken sentences. These devices often come with pre-built voices. And a lot of the time, those voices don鈥檛 match aspects of the user鈥檚 identity.

惭别苍诲辞锄补:听For example, the late physicist Stephen Hawking was British. But his electronic voice 鈥撀�

惭别苍诲辞锄补:听鈥撀爃ad an American accent.

Peng:聽But VocaliD can help people like Larry 鈥� who are at risk of losing their voice 鈥� create a digital clone of their voice. It鈥檚 a service known as 鈥渧oice banking.鈥�

Mendoza:聽To start with, you record yourself.

Mendoza:聽The VocaliD team then takes those recordings and the transcript, and feeds it to software that can "learn" to speak.

Peng:聽The technology isn鈥檛 perfect. But in short clips, you might not be able to tell the difference between these synthetic voices and actual human voices.聽

Mendoza:聽And building them has become much faster and cheaper. It鈥檚 pretty sci-fi.聽

[MUSIC]

Peng:聽Back in Canberra, when Alicia learned about VocaliD, she forwarded the article to Larry.聽

[MUSIC]

Peng:聽But Larry and Alicia also had hope.聽

[PAUSE]

笔别苍驳:听VocaliD delivered the synthetic voice to Alicia while Larry was still recovering from the surgery. She brought it to the hospital, and played his first words.

Mendoza:聽I love how Larry calls Alicia 鈥淟eesha.鈥�

Peng:聽Yeah, it鈥檚 so sweet.

Mendoza:聽Larry鈥檚 new voice software lives on his phone and iPad. For him and the people around him, hearing that voice has been pretty overwhelming.聽

Mendoza:聽And here鈥檚 Larry weighing in, speaking with his new digital voice.聽

Peng:聽So we actually sent Larry and Alicia some questions ahead of our interview. Larry typed out his answers on his VocaliD app, and Alicia played them for me.聽

惭别苍诲辞锄补:听For comparison, here鈥檚 what Larry鈥檚 old voice sounded like.聽

Mendoza:聽And here鈥檚 Larry鈥檚 voice clone.聽

Mendoza:聽Isn鈥檛 that amazing? When our producer Sam first heard Larry鈥檚 new voice she didn鈥檛 realize it was from an app. It sounded so authentic.聽

Peng:听驰别补丑!听

[PAUSE]

笔别苍驳:听At the time we reported this episode, Larry still had a lot he needed to work through. He was resting a lot and trying to build back his strength. He also didn鈥檛 lose the ability to speak entirely, and will be going through voice therapy. But he doesn鈥檛 know how much he鈥檒l be able to sound like his old self.聽

Mendoza:聽And then in terms of his new voice, needing to type each sentence before it鈥檚 spoken out loud slows down the conversation a lot.聽

Peng:聽Also Larry hasn鈥檛 used his voice app enough to know if it lets him express himself effectively in different interactions.

[PAUSE]

Peng:聽Still, Larry and Alicia are really thankful they found VocaliD when they did.聽

Mendoza:聽But Larry also said something that really stuck with both of us.

Mendoza:聽To me, that meant that although our voices are a big part of who we are, they鈥檙e not everything. Larry might never again speak in the way that he used to, but he can still hang on to and express important parts of him.聽

Peng:聽Right. In fact, Larry said he and Alicia still connect in many of the same ways they always have.聽

[MUSIC]

Peng:聽So Jess, this whole conversation makes me think of my experience, both as someone raised by parents who are disabled, and as a reporter who has covered disability. I know that people can be resilient, and they can find all kinds of ways to lead meaningful lives, and express themselves.

Mendoza:聽For sure. And even when they don鈥檛 have abilities that other people might see as essential, people can still be true to themselves and who they are. For our next story, we meet a young man who has had a speech disability since birth. And he鈥檚 getting a personalized voice for the first time.聽

Peng:聽It鈥檚 a fascinating process. It actually makes me think of Episode 1. We talked about how your voice is shaped by all these things: your upbringing, how others perceive you, your own self-image of course. And, in many ways, crafting a synthetic voice feels very similar. We鈥檒l be right back.聽

[MUSIC]

Clay Collins:聽Hi, I鈥檓 Clay Collins, an editor here at the Monitor. I hope you鈥檝e found a lot to take away from our podcast, which wraps up with this episode. Please rate and review us wherever you listened! And if you know someone you think would like Say That Again? as much as you have, please share your favorite episode 鈥� or send them to the full series at CSMonitor.com/SayThatAgain. As always, thank you for listening.

[MUSIC]

Mendoza:聽Welcome back. This is Say That Again? I鈥檓 Jess.

Peng:聽And I鈥檓 Jing.聽

惭别苍诲辞锄补:听Our next guests are the Boyles. They live in New Jersey. And we visited them in early March.聽

Peng:聽That鈥檚 Sean Boyle. He鈥檚 18 years old, although the voice he鈥檚 using might sound much older.聽

惭别苍诲辞锄补:听Sean is diagnosed with cerebral palsy. It means he has trouble controlling his movements. When we interviewed him, he was sitting in his wheelchair, with his mom, Jennifer, beside him. And he talked to us using a text-to-speech device called a Tobii. What you鈥檙e hearing is one of the Tobii鈥檚 pre-built voices.

Peng:聽The Tobii looks like a big, heavy-duty iPad. It鈥檚 mounted on Sean鈥檚 wheelchair, at about arm鈥檚 length from his face. The Tobii can read out pre-typed statements, like the ones you heard. It also allows Sean to string together sentences with his eyes, which the device tracks.聽

Mendoza:聽So the way it works is, Sean keeps his eyes on a word on the Tobii screen for one full second, and that tells the Tobii he's choosing that word. To start a sentence Sean might choose the word "I." And then, the Tobii will suggest verbs that can come next, like "want," or "need." And Sean does that until he completes the sentence.聽

Peng:聽The pace of the conversation was a lot slower than a typical interview. Also, for me, it was a bit hard to read Sean鈥檚 facial expressions. But there were times when his feelings came through quite powerfully.

[PAUSE]

Mendoza:聽We talked to the Boyles because they were working with VocaliD to build Sean鈥檚 first customized voice. The voice will replace the one he鈥檚 using now. And it blends Sean鈥檚 physical voice and the voice of his younger brother, Sam, who doesn鈥檛 have a speech disability, and whose voice is similar to Sean鈥檚.聽

Peng:聽It鈥檚 a complex process, and the technology is almost the least of it. There鈥檚 Sean鈥檚 sense of his identity, his family鈥檚 idea of how he 鈥渟hould鈥� sound, and sibling dynamics 鈥撀燼ll that comes into play.聽

Mendoza:聽But the hope, Jennifer says, is that the new voice will help Sean express more of himself when he鈥檚 interacting with others.

[MUSIC]

Peng:聽Sean needs care pretty much around the clock, so his parents were very present throughout the interview. They encouraged him to speak, and also interpreted sounds that Sean made in his physical voice.聽

Peng:聽That鈥檚 Sean鈥檚 dad, Neil.

惭别苍诲辞锄补:听Helping Sean express himself in as many ways as possible has always been important to his parents.聽

Peng:聽They got him his first Tobii when he was six. The device, like we said before, comes with a selection of voices, and they cover a range of genders, ages, and accents.

惭别苍诲辞锄补:听That voice is called 鈥淜enny.鈥� As Sean got older, they switched to a voice called 鈥淛osh.鈥�

Peng:聽This kind of situation is part of what inspired Rupal Patel to start VocaliD. Back in 2002, she was at a conference in Denmark. And she was struck when she saw a young woman and an older man talking to each other using the exact same synthetic voice.

笔别苍驳:听For people like Sean, who are born with a speech disability, the VocaliD team builds what it calls a 鈥渂espoke voice.鈥�

Mendoza:聽Jennifer Boyle found out about this service in 2016, at a summer camp for people who use synthetic voices. One of the campers, a girl named Maeve, had recently had a customized voice made.聽

Mendoza:聽Back then, Sean鈥檚 physical voice still sounded pretty close to the voice on his Tobii. But Sean began to outgrow the voice. So last year, Jennifer started looking into a customized voice for Sean. And so she turned to her younger son, Sam, whose voice was also starting to deepen.聽

Peng:聽But when Jennifer first asked Sam about it 鈥撀犅�

Peng:聽So we should note here: Sean and Sam have a pretty complicated relationship.聽

[PAUSE]

Mendoza:聽Eventually, Sam came around about donating his voice. And Sean liked the idea too. Both brothers wound up recording their voices for VocaliD.聽

Mendoza:聽Sam read out five hours鈥� worth of sentences, and Sean recorded various sounds that he can make.

Peng:聽And then VocaliD combined those two voices in what they call a 鈥渧oice smoothie.鈥� At the end of February, the first version of the voice came back. It sounded like this:聽

摆惭鲍厂滨颁闭听

Peng:聽You know Jess, it鈥檚 so fascinating that our perception of people鈥檚 voices is also wrapped up in what we think of their personality. Especially for people who are really close to us, I guess.聽

惭别苍诲辞锄补:听Yeah, and it makes sense that personality comes up a lot too when you鈥檙e building a synthetic voice. Here鈥檚 Dr. Patel again, from VocaliD.聽

Peng:聽Accents also come into play.聽

Mendoza:聽And as the technology develops, the voices get better, and more people use them, there are a lot of bigger ethical issues to consider too.聽

摆惭鲍厂滨颁闭听

Peng:聽For the Boyles, though, their main concern right now is just getting Sean the voice that feels right to him. When we visited them, they had just listened to the third iteration of the voice. Here鈥檚 what it sounds like.

Peng:聽Here鈥檚 another sample.聽

Peng:聽Jennifer compared these samples with the very first version.聽

Peng:聽Of course, the most important question is: What does Sean think?

Mendoza:聽When we brought this up, Jennifer suggested we communicate with Sean in a different way. The Tobii can be hard to use for more complex topics, and also it can get tiring for Sean.聽

笔别苍驳:听Right. So it鈥檚 actually pretty typical for the Boyles to use different ways to talk to Sean. In this case, Jennifer asked a yes or no question 鈥�

Mendoza:聽As Jennifer asked the question, she took each of Sean's hands. And then she gently shook each hand while saying an answer out loud. So the left hand was:

惭别苍诲辞锄补:听And the right hand:聽

惭别苍诲辞锄补:听Sean responded by dropping his right hand, so that his left was still up. Which meant that he did like it.

Mendoza:聽That鈥檚 Jennifer giving Sean options again.

Peng:聽And Sean鈥檚 answer:聽

Peng:聽Using the same method, Jennifer asked Sean what he thought about different features of the voice, like speed and clarity. It turned out the main thing Sean wanted to tweak was his pitch.

Mendoza:聽Finally, we asked Sean what he liked most about getting a personalized voice. Jennifer identified four things Sean was excited about.

Peng:聽Then she wrote these things down and showed the options to Sean.

[MUSIC]

Peng:聽Before we left, we got to share a very special moment with Sean. He really loves music, and loves it when his parents sing to him.

[The Boyle family is singing.]

Mendoza:聽Neil held him in his arms. And Sean had this big smile on his face as he sang along.聽

Mendoza:聽I just want to say, Jing, I was so touched by that experience. It just made me think again of what Larry said earlier. Our guest from Australia. Voice is important to who we are, but it's not everything. And after all these weeks where we鈥檙e trying to understand what our voices say about us, it just felt really meaningful to me to remember that everyone's experience is different. None of us are just one thing. We are how we sound 鈥� but we're more than that, too, you know?

Peng:聽Yeah, and you know, hearing Sean鈥檚 physical voice 鈥� I know that it could be hard to understand, especially for people who don鈥檛 know him well. But it鈥檚 still a voice that has value and power, and allows him to communicate his joy and his thoughts.聽

Mendoza:听驰别补丑.

摆惭鲍厂滨颁闭听

Mendoza:聽So Jing. This is it. Any final words?

Peng:聽I feel like over the past months, I鈥檝e learned so much about how we sound. And how we listen, and why that matters.

Mendoza:聽Of course! So clever.

Peng:聽But really I still feel like I know so little. And just hearing my own voice now. It鈥檚 this incredible thing that comes out of me, and helps me communicate, express different things. It does all these things for me. It鈥檚 really one of the biggest things in my life. But we rarely think of it that way. I mean, I think about it a lot more these days.

Mendoza:聽Has anything changed about the way you think about your voice, or even how we sound in general?

Peng:聽Yeah! I've become more sensitive to times when people express that they don鈥檛 like the way someone sounds, because sometimes you鈥檙e really talking about more than the way they sound. And for the listener, it鈥檚 a very impactful thing to hear. Because it can very easily be or feel like a judgment of your person.聽

Mendoza:聽I actually feel like it鈥檚 happening more often that I catch myself in situations where I can feel myself about to make a judgment about the way someone is talking. And then I feel like I have to hit a pause button in my brain. And I鈥檓 asking myself first, 鈥淥kay, Jess, can you understand them?鈥� And if the answer is yes, then what does it matter how they鈥檙e saying these words?

Peng:聽It鈥檚 a great button to have.聽

Mendoza:聽And I鈥檓 hopeful it鈥檚 a button that sticks around, that I make a habit of it. Because then it will feel like I鈥檓 moving forward from this project having learned something. But beyond that, Jing, I鈥檝e really enjoyed this journey, it鈥檚 been a lot of self-reflection. Some of it has been pretty heavy, but it鈥檚 been such a great adventure doing this project with you, Jing!

Peng:聽Oh my god. Remote hugs to you, Jess. And to all of you out there 鈥� we definitely still want to hear from you. If you have any comments, questions, or stories you want to share, you can send us a note at聽podcast@csmonitor.com. We really want to keep talking about it.

Mendoza:聽Season 2鈥� Season 2鈥�.

Peng:聽I hope so!聽

Mendoza:聽But a break first, I think.聽

摆惭鲍厂滨颁闭听

Mendoza:聽So one last time: This podcast was written, reported and produced by me, Jessica Mendoza.

Peng:聽And me, Jingnan Peng.

Mendoza:聽Our scripts were edited by Clay Collins, Trudy Palmer, and Samantha Laine Perfas. Sam also provided production support. Our sensitivity reader for this episode was Laken Brooks.聽

Peng:聽Our logo was designed by Jacob Turcotte. The artist behind the incredible illustrations on our site is Jules Struck. You can check out her work at csmonitor.com/saythatagain. And sound design for the series was by Morgan Anderson and Noel Flatt.

Mendoza:聽This podcast was produced by 海角大神, copyright 2022.聽

[END]