Digital actors Bashny.Net

Uncanny valley h4> Three years later, ironically, Tom Hanks himself became a "source code" for a character in a feature-length animated film Polar Express i> (The Polar Express):



Some eerie feeling from those of all the characters, right? This so-called «зловещая Valley », psychological effect, in which the spectators a feeling of discomfort (anxiety, fear, neuyutnosti) contemplation of the characters that appear almost as a living person. This unpleasant undercurrent of feeling occurs when a character loses all features and simplified caricature that explicitly say our consciousness that it is a fake, imitation. And it was, oddly enough, one of the main challenges to the creation of advanced digital actors.

In the mid-2000s, it seemed that we would never be able to achieve such realism artificial people to get rid Finally, the effect Uncanny valley. However, in 2008 leaves a wonderful film The Curious Case of Benjamin Button i> (The Curious Case of Benjamin Button). Efforts animators Brad Pitt for the movie received image and an old man and a teenager, while remaining recognizable antiquity Brad.



And now, six years later, in most scenes it seems that it's just a very well made-up actor. But it is not so:



An example of this film showed that in the episodes already becomes possible to get rid of the sinister effect of the valley, as the cost and complexity of do not yet allow full double. For example, in some scenes of the film Iron Man 3 i> (Iron Man 3) has been activated digital copy actor Guy Pearce, because he could grow a beard to shoot another movie.

Someone even very high-quality simulation of the monkeys in the Dawn of the planet of apes (Dawn of the Planet of the Apes) is not only surprising, but some inexplicable alertness. Although it is possible to explain the unusual humanized monkey persons.



The disadvantage of the effect of "Uncanny valley" for cinema industry is that instead of surprise and delight the viewer can experience negative emotions. And this is a direct path to the commercial failure of the film, and why then spend the time, effort and a lot of money in creating special effects? And in today's blockbusters is the main item of expenditure of all kinds of resources, not just financial.

How many resources are now required to realistically simulate a human? Of course, this is highly dependent on a number of conditions, but on average the hospital every frame requires about 30 hours of work a large number of very powerful servers. A five-minute film of 7200 frames at 24 frames / sec. At the time, the real breakthrough was the use of technology motion capture, without which it is impossible to achieve complete realism movements of 3D-models.

Digital Ira h4> year and a half ago, Nvidia released a demo video rendering technology persons Digital Ira.



It is very realistic depicts early polysevshy man of thirty with a small speaking a little on the nose. When viewed from a distance or in passing, that of a living person does not tell the difference. Even can be confused if you do not see the eyes, despite the fact that for a realistic simulation of their movements are used serious computing power.

Surprisingly, but still eye digital actors could not "give in" to the animators, they do not get completely indistinguishable. Here necessarily reflect on the meaning of the phrase "Eyes - the mirror of the soul" ... Every feature of this digital model requires serious computing: folds and irregularities of the skin, blood vessels, the movement of facial muscles. A man can express pleasure, confusion, anger, surprise, boredom, pleasure and anxiety. And almost at the same rate as that of a living person. This is a very important factor, since our brain immediately says is too low or high rate of change of facial expression, resulting in us a feeling of falsity.

How does it work: first scanned human head model, digitized movement of facial muscles, jaw, eyes. In addition to the mechanics of calculated reflected light sources, shadows card with a particular facial expression, while talking.

So far, it has been demonstrated in two versions Digital Ira: from Nvidia in March 2013 at a conference on technology GPU, and from Activision in July 2013 at Siggraph. Nvidia launches it on one of their cards, while still performing 4, 9 trillion operations per second. Activision working version on the PlayStation 4 GPU capacity of 1, 84 teraflops.

Movies h4> As you can see, we are already close to the main goal, which is the special effects industry has set itself - creating interactive photo-realistic digital actors. And then there are a number of questions and doubts.
Imagine that developers finally managed to create digital humans that are indistinguishable from the real thing (and it will happen sooner or later). This completely overturn our understanding of acting skills and the role of actors in general. Games turn into a kind of interactive movies. Believe TV news in general will become meaningless, given the possibility of forgery of any movie. Although now they can not be trusted when it comes frankly.
Despite the technical ability to create almost entirely realistic digital actor, cost and complexity of such that it is cheaper to hire highly paid Hollywood star. The main reason - the need to use very large computing power, which are constantly becoming cheaper. And when the cost of creating the perfect, perfect "digicam" will be comparable to or lower than hiring a good actor, that's when we expect rapid changes.

Artificial litsedeev will be no need to divide by the dramatic, comedic, etc. They will be able to play any role of any character. Almost all films with actors using digital three-dimensional models of living of people in the same "The Curious Case of Benjamin Button." However, the main purpose of experts in special effects is to create arbitrary characters that are indistinguishable from humans. Arbitrary age, appearance, gender, height, completeness, etc. Although digital copies will also be in demand, imagine how many wonderful, great actors could be immortalized in forever young "digital camera", using them for decades. While we all grow old, the same age the actors are no longer able to fulfill many roles.

Games h4> It is possible that we are waiting for the flowering of "filmoigr" with lots of forks scene in which the story will develop in one of several scenarios, depending on the make "igrozritelem" choice. This principle is already used in games, one of the examples - two games of the series Walking Dead .





Here players show some story that it looks like a spectator. But occasionally there are situations that require the player interactivity and choice, as do the character. Depending on this story can develop in a different direction, but in certain boundaries.

This approach will be to develop further, creating a small world with a lot of stories. Players will be able to choose the characters, whose storyline they want "igrosmotret", with the ability to switch to another character if his story seems interesting.

The logical step h4> Today in movies and games are usually created characters often with the help of technology motion capture. However, until cinema games yet far by the degree of certainty. On the other hand, is a matter of budget and computing power: a movie viewer shows the final rendered image, and in games all the work falls on the long-suffering computers and consoles users.



Yet to call actor a digital copy of a living person can not. After all, the actor itself forms its image, behavior, emotions, gestures. And after we will be able to play with an acceptable cost a person's appearance, the next step will be the creation of technology "aktёrstva" digital characters. That they are not repeated sequence of actions laid down by the operator, not "lost" pre-scanned mimic diagram. In fact, the digital actor to be in some way artificial intelligence that will choose for myself as him to play a role or a scene, how to express emotions. To act on the same principle that we, when we choose how we ourselves stories in a given situation, how to show their feelings and thoughts.

Of course, the dream is not harmful. The creation of such technologies will require addressing the challenges that currently stand in the way of development of artificial intelligence. And even further, because the intellect must be able not only to "speak", but also "express themselves" non-verbally.

The problem is compounded by the fact that we still need to understand yourself how the acting talent, what psychological (or physiological?) Make some people especially the actors? We also do not know the mechanism by which the manifestation of some emotions can affect our facial expression, posture, gestures. Scientists are still arguing that primarily from the standpoint of the underlying psychophysiological mechanisms - whether we feel happy, smile because, whether we smile because it felt happiness.

So it will take more than a decade until we can talk about the possibility of creating a full-fledged digital actors. Until then, we'll "settle for" digital double varying degrees of credibility. By the way, it is believed that now we need to start to fix the detailed appearance, facial expressions and behavior of various politicians, scientists, influential people - all those whose digital copies could in the future be used to create movies and games. For the sake of historical accuracy, so to speak. In fact, it is no different from a permanent photography and video, with the use of these materials for a variety of tasks. There is though, and this opinion is that digital actors are evil, and the real art of acting is available only living people. After all, is it artificial intelligence can compete with the brilliant talents of actors past and present?

Source: habrahabr.ru/company/asus/blog/245079/