GAILLAC.- Guilhem Gallart used to speak with a thick, southern French accent, his voice deep and slightly nasal, topped by a faint lisp.

Then, in 2015, he was diagnosed with amyotrophic lateral sclerosis, or ALS, an incurable neurological disease that slowly paralyzed his muscles from head to toe, leaving him bedridden and forcing him to use a voice-synthesizing computer program to speak.

Now his family jokes with him that he sounds like a GPS device. His wife and two daughters, Gallart said, sometimes call his old cellphone number just to hear his voicemail greeting.

Losing his distinctive voice, he said, has felt like surrendering an essential part of himself, as sound has been his life’s passion. Better known as Pone, he is a music producer and beatmaker who once belonged to one of France’s most popular old-school rap groups, the Fonky Family.

In a bid to recapture his signature vocal sound, Pone, 50, has embarked on a slightly quixotic and still-unfinished quest. Because there were not enough old recordings of his voice to feed into a computer and create a synthetic replacement, he asked a comedian to record an imitation of what he used to sound like — and used that as a basis instead.

“A voice is a very personal thing,” Pone said, propped up in bed at his home in Gaillac, a small town near Toulouse in southwestern France. The way you talk “says volumes about your character,” he added, instantly conveying so much of who you are — even if, as he jokingly described his voice in a documentary about his life, he “really sounded like an idiot.”

In 2020, he contacted Marc-Antoine Le Bret, a 37-year-old comic known for celebrity impressions, asking for his help in creating a new voice to match his old one.

“The challenge was absolutely crazy,” Le Bret recalled thinking. He had never met Pone but immediately agreed. “It’s the most beautiful challenge of my lifetime.”

Le Bret immersed himself in old audio and video recordings of Pone, like clips from a family vacation to Morocco. He also watched footage of Pone’s band.

Pone did not sing or rap in the Fonky Family, which sold hundreds of thousands of albums in the 1990s and 2000s with fiery raps about the hardscrabble streets of Marseille. But Le Bret listened for snatches of Pone’s voice in interviews, behind-the-scenes footage of concerts and studio chatter in recording sessions.

Le Bret next worked on refining his impression. But it wasn’t easy.

Pone was born and raised in Toulouse, where he fell in love with hip-hop as a teenager, before moving to Marseille. The two cities have slightly different but equally strong accents, with the silent “e” not always so silent. (The pronunciation of “Pone” varies from Paris to the south, but the “e” is silent in both locales.)

Le Bret had to capture Pone’s tone, too, and said he had “tried to absorb” Pone’s way of expressing himself, without slipping into caricature. “He’s a wisecracker, he loves irony, and that was important,” he said.

Getting the intonation right was essential to Wahiba Gallart, Pone’s wife, and the last person who could understand him before he switched to the impersonal, digitally generated voice.

“Sometimes he’ll tell us something, and the computer’s tone is flat, cold,” she said. That makes him sound more abrupt than intended. “It was important for him to recover as much of his former self as possible, for our daughters,” she added.

Once Le Bret felt his impression was close enough, he voiced about 250 stock phrases in a recording studio, many of them nonsensical sentences designed to provide the right balance of phonemes to feed a program designed by CandyVoice, a digital voice processing and synthesizing company.

Gallart attended the three-hour session, correcting and guiding Le Bret. Pone and his family could hardly wait for the results.

Family is Pone’s pillar. His parents and half-siblings often visit the home in Gaillac, where wedding and vacation shots hang not far from gold and platinum records celebrating the rap and R&B hits he produced in his solo career after the Fonky Family broke up in 2007.

His daughters — Naïla, 15, and Jasmine, 12 — often clamber onto his bed to chat, watch movies or play video games.




And last year, the daughters’ smiles grew wide as they gathered to listen to Pone use the new voice for the first time — a moment captured by a national television news crew.

The results earned a mixed review.

The voice was synthetic but distinctly southern and very close to his original one. Unfortunately, it was also buggy. Longer sentences were garbled; syllables were occasionally dropped. Pone was quick to spot the glitches.

“I was dealing with a sound engineer, an expert,” said Jean-Luc Crebouw, the founder of CandyVoice. He said Pone had grown used to the neutral-sounding synthetic voice, which made the kinks in the new one stand out even more.

But the experiment was successful enough that Pone said he wanted to keep going, and he has contacted other companies to fine-tune the new voice.

“Using my actual voice was a plus,” he said. “I was able to express myself again.”

New artificial intelligence and computer-generated speech tools are offering hope to patients with speech-impeding diseases, although these approaches typically require people to either still be able to speak or have a fairly large library of recordings of their voice.

Pone, who lost his voice before these technologies were available, needed the more creative solution of partnering with Le Bret. In recent weeks, he has tried different AI tools, but the results were “very disappointing,” he said, and for now, he has reverted to his old synthetic voice.

When he wants to speak, an infrared sensor tracks his eyes as they move across a keyboard on a screen, enabling him to write text that is then read out in a monotone. The process is slow. During an interview, minutes of silence between each answer were broken only by the rhythmic whooshing of his artificial respirator and the occasional beep of his gastric feeding machine.

The challenge Pone set himself to recapture his voice is not the only one he has taken on since being diagnosed eight years ago with a disease whose mean survival time, according to the ALS Association, is two to five years.

Pone can no longer spend hours improvising with samples in a studio. But since his diagnosis, he has started a music label, and using his eyes to operate music software, he has released several albums, including one based entirely on Kate Bush samples. He also produced a mix for the handover ceremony of the Paralympic Games, held in Tokyo in 2020 and scheduled for Paris in 2024.

He started an “ALS for Dummies” website and wrote a recently published autobiography. Pone said he had always been hyperactive. “Now,” he added, “I’m just in more of a hurry.”

Since the diagnosis, Pone said he had become a “better version” of himself, uncluttered by fleeting distractions and focused on the essentials.

ALS “left me what mattered most,” he writes in his autobiography. “My mind and my heart.”

He converted to Islam in the mid-2000s, and he said his faith had helped him find inner peace. “It’s all about acceptance,” he said. “And I have accepted.”

He does miss life’s tactile pleasures, like walking on a sandy beach or hugging a loved one. And landing a punchline is hard when you have to type it with your eyes.

While the disease has stolen his speech and much else, he and his family said it had only strengthened their bonds.

Jasmine is too young to remember what her father sounded like and said she was excited to hear a better version of his “new” voice. Still, the current one has grown on her. “When he is joking or is happy, I almost feel like the intonation changes,” she said.

“It’s not really the voice of a robot anymore,” she added. “It’s my dad’s.”

This article originally appeared in The New York Times.