What if meeting Dalida, Jean Gabin or Lady Di were possible in 2022? We would dream of it, Thierry Ardisson has done it (almost). In Hotel del Temps, the first edition of which will be broadcast on Monday 2 May on France 3, the star presenter interviews dead actors, singers and politicians. How is that possible? Through one of the applications of artificial intelligence: deepfakes. More precisely, the presenter and his team relied on the technologies of two French companies, Ircam Amplify and Mac Guff, to recreate, with one, the voices of the stars that disappeared, and with the other, their faces. .
Behind the algorithmic experiment, two French companies
“One of our areas of work is voice in the broadest sense,” says Nicolas Pingnelain, commercial director of Ircam Amplify. Created in 2020, the company aims to financially promote the fundamental research carried out at the Institute for Research and Coordination of Acoustics / Music (Ircam). “Up, this means detecting emotions, markers of stress or anxiety, downstream, synthesizing a voice capable of reproducing this type of variation. The human being is even more sensitive to the voice than to the image, he continues, a little allows us to detect if it is false. No wonder Thierry Ardisson’s team turned to French researchers: a monotonous voice like that of our smart assistants to take on Dalida’s intonations would never have worked.
It was also necessary for the faces to create the illusion of reality. The team turned to Mac Guff, a 35-year-old French studio, creator, among others, of Me, ugly and bad. “We have always had a research and development dimension, which allows us to stay ahead,” said its president, Rodolphe Chabrier. So we were the first to do morphing, relief … ”Five years ago, the entrepreneur embarked on artificial intelligence. Mac Guff soon built two tools: the talking image, to “talk” still images, and the face engine, dedicated to recreating 3D faces. Both built from deep learning models, these technologies made it possible to model the faces of stars from the past on the bodies of the actresses and actors who gave the answer to Thierry Ardisson during the filming of the docu-fiction.
How do we resurrect the dead?
Because Thierry Ardisson is interested in: Hotel del Temps it is a documentary work. And it is true: to train the algorithmic machines that reconstruct faces and voices, it is necessary to equip them. These training datasets consisted of all possible files of each of the celebrities that the program brings to life. “Appearances on television, images donated by families, films in the case of actors like Jean Gabin, lists Rodolphe Chabrier … We take everything we find! “For the voice, it’s a little simpler,” explains Nicolas Pingnelain. “The important thing is to have samples of all the phonemes, all the syllables in the language. Then, with a few tens of minutes of speech, we can have a good result. “As for the words spoken by Dalida, Lady Di or Coluche on the show,” said Thierry Ardisson, “they were all really spoken during the life of celebrities.
The result works pretty well … But in any case, we can’t help but feel uncomfortable. Could it be the Valley of the Uncanny effect, which makes robots trying to imitate humans a bit unpleasant to the person watching them? Is it the fear of a certain smoothing of the faces, when one dares in the nose a little too fine of the artificial Dalida, in the photos of shootings? “Just two weeks ago we had Dalida’s voice validated by her brother Orlando and he himself can’t distinguish between the synthesized and the real version,” says Nicolas Pingnelain. “When they saw the result, Gabin’s children found their father’s expressions, they were amazed! adds Rodolphe Chabrier. The production also made the decision to “magnify” the stars, he says, “to look for the ‘iconic’ Dalida, as Thierry says.” In this case, the discomfort may come from ethical issues: is it acceptable to bring the dead back to life? Response from all who have worked there hotel of the time : not only is there no law banning it, but the rights holders have reviewed everything, the scripts, the faces, the voices and everything has been validated.
“A tool has no ideology, emphasizes Thierry Ardisson, this point is very important in the face of any a priori.” Rodolphe Chabrier conveys the same message and we can not help but smile: in technology, the argument of neutrality is a classic. In the world of artificial intelligence, however, more and more researchers and activists believe that posing the problem in terms of good or bad use is not always relevant: in some cases, one would simply have to refrain from developing a a tool. Per hotel of the timeThat being said, it is clear that Thierry Ardisson has been following an idea that has haunted him for a long time: in 1994, he “resurrected” John Lennon in Gone are the days. In 2002, Victor Hugo was invited to the set ofWe will have seen it all, alongside Renaud and Guy Bedos, alive and well. Replacing similar ones with deepfakes, maybe it’s just keeping up with the times?
This would lose all the dimension of innovation required to create the program: Rodolphe Chabrier and Nicolas Pingnelain emphasize the innovative aspect of the program and the work it required. “60 minutes of recomposed faces more than 90 minutes of program, this had never been done before! exclaims Mac Guff leader, if only because accessible technologies before the company built its own algorithmic models would have made the operation too expensive. For speech synthesis, “our researchers modified the tool less than a month ago, and will probably do so again after the release ofhotel of the time says Nicolas Pingnelain: the show used applications directly from basic research. Every scientific advance will improve the programs that have been built for the occasion.
A double cultural opportunity
In fact, maybehotel of the time it opens up two cultural axes at once: the ability to get to know the interviewed stars better, first. After all, Thierry Ardisson’s stated goal is to “spectacularize culture and make knowledge captivating.” It is also an opportunity to see what even a young technology can give in the field of visual and sound special effects. Deep fakes have already been used in series, by Mac Guff, to rejuvenate Mathieu Amalric in The office of legendsor by the Russian operator Megafon, to play Bruce Willis in an ad without him actually acting.
Hard to predict yes hotel of the time it will mark a turning point in itself. On the other hand, working on this particular case allowed Mac Guff developers and Ircam researchers to move forward in their respective fields. And among the applications they imagine for their new algorithmic tools, Rodolphe Chabrier and Nicolas Pingnelain outline such varied avenues as voice authentication (to finally abandon passwords), dubbing (no more French voices that do not correspond to northern actors -American?) or, more classically. , retouching faces and then bodies. “With the Face Engine, you can recreate, rejuvenate, age, grow, transform someone into a burn victim in a very realistic way, for film or video games,” illustrates Rodolphe Chabrier. Who knows what show Thierry Ardisson will invent when you can also play about bodies and environments.