When Cynthasia launched in 2017, its main purpose was to meet the AI version of real human faces – for example, former footballer David BeckhamDifferent with sounds speaking in different languages. A few years later, in 2020, he began to give companies the opportunity to make professional -level presentation videos to companies signing up for its services, which included AI version or consent actors of staff members. But the technology was not perfect. The physical movements of the avatar can be nasty and unnatural, their accents sometimes slip, and the emotions indicated by their voices are not always similar to their facial expressions.
Now the synthetic avatar has been updated along with more natural procedures and movements, as well as expressing sounds that protect the speaker’s accent better. For the corporate clients of Cynthia, this avatar will make a make of financial results, internal communication, or staff training videos.
I have shown the video just as baseless by displaying my avatar because it is technically impressive. It is quite smart to pass as a high definition recording of a cherry corporate speech, and if you didn’t know me, you probably think that was exactly. This demonstration shows how difficult it is to distinguish artificially. And long ago, these avatars will be able to talk to us too. But how much can they get better? And what can a dialogue with AI clone do with us?
The process of creation
When my ex -colleague Melissa visited the London Studio in Santhaisia last year to make him an avatar, he had to go through a long process of eliminating the system, reading the script in various emotional conditions, and eliminating the sounds needed to create its avatar sounds and strengths. Since I was standing in the bright room after 15 months, I was relieved to hear that the process of creation was significantly smooth. Sanathisia’s technical supervisor, Josh Baker Mendoza, encourages me to move indicators and my hands during natural conversations, while at the same time warn me not to move too much. I have repeatedly repeated a shining script that is designed to encourage me to speak emotionally and enthusiastically. The result is slightly as Steve Jobs has been raised as a blonde British woman with a low, nauseous voice.
It also has the unfortunate effect that they are giving me a voice like an employee of Sanathisia. “I am happy to show you what we are working on today with you today.” So be ready to be part of something that will force you to go, ‘wow!’ This opportunity is not just big – it’s memorable.
Just an hour later, the team has all the footage it needs. After a few weeks I received two avatars of myself: one previous Express 1 model -driven and the other was made with the latest Express -2 technology. The latter, Cynthisia, claims that it makes its artificial humans more and more to the truth with the people on whom they are modeling, which are complete with more expressive gestures, facial movements and speech. You can see the results for yourself.
Courtesy Recipes
Last year, Melissa found that her Express 1 -driven avatar failed to meet her translantic accent. His emotions were also limited – when he asked his avatar angrily to read the script, it seemed more dirty than anger. In the subsequent months, Cynthysia has improved the Express -1, but my avatar version is made of the same technology and is still struggling to integrate physical movement with speech.
On the contrary, I am surprised that my new Express -2 avatar looks like me: his facial features are my own right mirror. His voice is also strictly correct, and although it indicates more than me, his hand movements usually marry what I am saying.