In online video games such as “Fortnite,” players can pay to customize their on-screen appearances. Now a Cambridge startup called Modulate wants to let players customize their voices, too, so an adolescent newbie can sound like a grizzled gaming veteran, or female players can come across as men.
There are already products on the market, including Voxal and VoiceMod, that electronically alter human voices. Cofounders Mike Pappas and Carter Huffman say Modulate will produce better results because it uses artificial intelligence to analyze recordings of human speech, such as words spoken by a professional actor. The software measures the tone, pitch, and emotional intensity of the speaker’s voice, then applies these qualities to any user’s speech in real time.
“It’s basically swapping out your vocal cords digitally,” Pappas said.
Modulate has just raised $2 million from angel investors and a couple of venture capital firms, 2Enable Partners and Hyperplane Venture Capital. Now it’s talking to video game companies about incorporating its voice technology.
Titles like Epic Games’ “Fortnite” cost nothing to play, but the company makes millions of dollars by selling “skins,” customized digital bodies that players use inside the game. Because up to 100 people play in each round, buying a skin is a good way to make a gamer stand out from the crowd.
Modulate plans to team up with video game companies to sell “voice skins,” custom voices to go along with the players’ custom bodies. For instance, last November “Fortnite” briefly let players buy body skins that matched the appearance of NFL football players like Tom Brady of the New England Patriots. With Modulate on board, Epic could have also sold something like Brady’s voice.
“There are 200 million monthly players of ‘Fortnite’ alone,” said Mike Dornbrook, former chief operating officer at music game company Harmonix Music Systems and an early investor in Modulate. “You don’t need that many paying customers to have a pretty good business.”
And of course there are many more free-to-play games, such as the recently released “Apex Legends,” where Modulate might find a ready market for its voice skins.
Pappas and Huffman foresee a host of other uses for Modulate software. For instance, Hollywood studios might use it to re-record dialog without having to pay stars for additional work. Or music recording companies might someday produce recordings featuring the simulated voices of famous singers.
Pappas and Huffman became buddies while earning undergraduate degrees in physics from the Massachusetts Institute of Technology in 2014. Rather than face an uncertain future as research physicists, they went looking for a business they could launch.
While Huffman was working on video processing software at the Jet Propulsion Laboratory in California, he realized that the same methods could automatically transform human voices. When he discussed the concept with Pappas, the idea for Modulate was born.
The system has its limitations. Voices have a clipped, slightly inhuman tone, though Pappas and Huffman said that the quality will improve as they continue to polish the software. Also, Modulate imitates only the tone of the voice, not a person’s unique vocal quirks.
The technology might also lend itself to abuse. Imagine a con artist phoning you to ask for money, in the voice of your college-age daughter. But Pappas and Huffman say they plan to limit distribution of their software to ethically responsible companies. Also, the audio produced by the Modulate system will carry digital “watermarks” that can be detected by a computer, as proof that the voices came from a computer and not a real person.Hiawatha Bray can be reached at firstname.lastname@example.org. Follow him on Twitter @GlobeTechLab.