[ad_1]
Fred, trans man clicked the mouse and his sad tones suddenly sank deeper. He had included voice change algorithms that provided what sounded like an immediate vocal cord transplant. “It’s Seth,” he said of a person testing a Zoom interview with a reporter. He then switched to speaking as “Joe,” whose voice was more nasal and optimistic.
Fred Jane’s friend, a trans woman who also tests the prototype software, smiled and showed some artificial voices that she liked for their feminine sound. “This is Courtney” – bright and optimistic. “Here’s Maya” – higher, sometimes too much. “This is Alicia, the one I find has the most vocal variation,” she concluded softer. The problems were minor enough to provoke the fleeting thought that the couple might not have joined the conversation with their “real” voices to begin with.
Fred and Jane are early testers of technology from the startup Modulate, which could add new fun, protection and complication to online communication. WIRED does not use its real names to protect its privacy; Trance people are often bullied online. Software is the latest example of the complex potential of artificial intelligence technology that can synthesize real-world video or audio, sometimes called deep fakes.
Modulate co-founders Mike Papas and Carter Huffman initially thought that the technology they called “voice skins” could make games more fun by allowing players to take on the characters’ voices. As the couple offered studios and recruited early examiners, they also heard a chorus of interest in using voice skins as a privacy shield. More than 100 people asked if the technology could alleviate the dysphoria caused by a mismatch between their voice and gender identity.
“We’ve learned that a lot of people don’t think they can participate in online communities because their voice puts them at greater risk,” said Papas, Modulate’s chief executive. The company is now working with gaming companies to provide voice skins in ways that offer both entertainment and privacy, while promising not to turn them into a tool for fraud or harassment.
Games like Fortnite and social apps like Discord have made it common to join voice chats with strangers on the Internet. As in the early days of texting over the Internet, the voice boom unlocked both new pleasures and horrors.
The Anti-Defamation League found last year that nearly half of gamers experienced harassment through voice chat while playing, rather than through text. A sexist stripe in gaming culture makes women and LGBTQ people wanted for particular abuse. When Riot Games starts team shooting Evaluation in 2020, executive producer Anna Donlon said she was amazed to see a culture of sexist harassment emerge rapidly. “I don’t use voice chat if I go in alone,” she told WIRED.
Modulate’s technology is not yet widely available, but Pappas says it is in talks with gaming companies interested in implementing it. One possible approach is to create modes within a game or community where everyone is assigned a voice skin to match their character, whether they are a rough troll or a knight in armor; alternatively, the votes can be set at random.
In June, two of Modulate’s voices launched a preview of an app called Animaze, which turns a user into a digital avatar in live streams or video calls. The developer, Holotech Studios, offers voices both as a privacy feature and as a way to “transform your voice to better fit a character of a different age, gender or body type than your own.” Modulate also offers gaming company software that automatically notifies moderators of signs of abuse in voice chats.
Modulate’s voice skins are powered by machine learning algorithms that adjust the audio patterns of a person’s voice to make them sound like someone else. To teach its technology to sound many different tones and timbres, the company collects and analyzes audio from hundreds of actors who read scripts created to provide a wide range of intonations and emotions. Individual voice skins are created through tuning algorithms that reproduce the sound of a particular voice actor.
[ad_2]
Source link