Before the movie "Alita: Battle Angel" was finalized, Fox would never have imagined that there would be a science fiction film called "The Wandering Earth" mademade IN CHINA to steal the limelight. In my eyes, "The Wandering Earth" is like a toddler, both new and imitative, both aura and clumsiness. The kid who has just walked into the science fiction film club does bring new faces and new air, but he still has a long way to learn. "Alita: Battle Angel" is another exquisite product bred by the science fiction culture of the United States and Japan and created by the Hollywood film industry. What we just did was 0 to 1, and Hollywood was doing 100 to 101. Although there is no need to worship and copy everything, it is always right to learn with an open mind.
Alita: Battle Angel Mainland Movie Poster
From the perspective of film culture, Alita: Battle Angel has too many Elements of Hollywood Sci-Fi that we are familiar with:
First, cyberpunk: The steely city in the film, similar to the Saka star in Blade Runner (1982) and Thor 3 (2017).
Second, Mech Warrior: The favorite thing of Japanese science fiction culture, the previous Hollywood movies based on it seem to be less successful, such as "Pacific Rim" (2013), "Ghost in the Shell" (2017), compared to "Alita: Battle Angel" completely Hollywood.
Third, wasteland movies: human civilization rebuilt after the Cataclysm, similar to films such as "The Postman" (1997) and "Mad Max" series.
Fourth, film noir: In addition to "Alita: Battle Angel" director Rodriguez's proud "Sin City" (2005), there is a lesser-known "Soul Shifter" (1998), and even the old version of the four "Batman" is also film noir style.
Fifth, dystopia: individuals single out totalitarian societies, such as "The End of Tearing" (2002) and "Elysium Space" (2013).
……
The worlds of Wasteland and Cybertron in Alita: Battle Angel
Alita: Battle Angel is a highly integrated packaged and upgraded version of these film elements. As for why this movie has been so dismal at the North American box office, I can't figure it out. Maybe the market publicity is not in place, or maybe the Western audience has aesthetic fatigue with these things, and maybe the CG character of Alita in the film accidentally stepped into the "Weird Valley" (the "Weird Valley" will be explained in detail below).
My biggest point of emotion is that this is really not a 0 to 1 movie. The birth of this film is not only behind the Hollywood film industry with strong capital and first-class technology, but more importantly, Hollywood has accumulated enough film culture and is still inheriting and developing. Further, behind the film culture, it is inseparable from the historical background of industrial civilization experienced by the United States and Japan. For an agrarian civilization, the bicycle may be their science fiction; for a country full of bicycles, the car may be their science fiction; and for a country full of cars, the Transformers are their science fiction. Without two industrial revolutions, there would be no sci-fi cultures like cyberpunk and steampunk; without the Cold War and nuclear crisis, there would probably be no such imagination as film noir and wasteland movie.
The above is all about the dragon of Alita: Battle Angel, and the following only talks about the CG character of Alita, who is the eye of this dragon. Regardless of the film's eventual box office, Alita will be a landmark CG character who represents a new level of cinematic technology. Limited by space, the CG characters mentioned here are limited to characters such as humans or orangutans close to humans, and CG characters such as dinosaurs and monsters are not included. Before entering the main text, you must understand the following three points:
First, creating a CG character that can be accepted and recognized by the audience is not only a technical problem, but also involves the aesthetic psychological level. For example, the audience's recognition of monsters will not be too high, and anything can be like a monster; it is difficult to make a living dinosaur, because what a dinosaur looks like is a little bit in everyone's mind; it is more difficult to be a CG mouse than to be a CG dinosaur, because the realism of fur was once a big problem in the CG industry.
Second, for CG characters, the intensity of identity that needs to be established in different film contexts is also different. For example, Avatar's Na'vi and Alita, the former because they are aliens in a virtual alien environment, so they only need to do the appearance of people; while Alita's story takes place in the real world, which needs to establish a higher sense of identity for the audience, so the requirements for technology are higher. In fact, Alita is definitely the most technical screen CG character at present, not one of them.
Third, our people's recognition of faces is surprisingly high. This is the fundamental reason why the more CG characters resemble real people, the harder it is to do. For a CG person, as long as there is a slight slight difference from the real person, we can feel it. And once you have this feeling, it is likely that you will have to "play" when you watch a movie.
Alita, a CG character who strives to achieve a live-action effect
Let's comb through how the CG characters on the screen evolved to Alita step by step.
1. Flat Knight - the first CG character in history
Movie poster for The Young Sherlock Holmes (1985).
This "Young Sherlock Holmes" is a decent detective mystery film, produced by Spielberg, and George Lucas's Industrial Light and Magic Company is responsible for the special effects. The film went down in history for creating the first CG character in history.
This CG character appears around the 22nd minute of the film. A priest hallucinated after being hit by a poisoned arrow and saw a painted knight on the window jump down and approach him step by step with a sword. This clip is only a few seconds, and if viewed separately, there are five shots in total. Saying that this is a CG character is a little reluctant, not only because it is a two-dimensional flat thing, but also from beginning to end, it is a walk and a sword-raising action, not to mention the change of facial expression, the whole face is drawn and does not move at all. The most difficult part of this segment in production is in the fourth shot, when the knight walks forward, the camera rotates from his front to his back, which is the most technical of the five shots.
Five shots of CG knights in Young Sherlock Holmes
2, Gollum - motion capture technology to create "human-like" characters
The landmark CG character Gollum
In 2001, there was a blockbuster film that used motion capture technology, and then won the Oscar for best supporting actor, which is the famous "Lord of the Rings". In fact, the strength of The Weta Studio, which was dominated by traditional models and props, could not be compared with Industrial Light and Magic, but the superb performance of actor Andy Serkis prompted the director to decide to use the technology in order to bring the magical role of Gollum to the screen. Since there was no very perfect real-time technology for motion capture and live-action interactive performances, Gollum's shooting was divided into four parts. The first part is where Andy Serkis performs with the actors in white, and the second part is a computer in which Serkis is wiped out and replaced with a gollum. The third part is Serkis performing alone, wearing a capture suit with motion points, and finally loading facial performance data onto Gollum's CG model.
The first generation of motion capture technology used by Weta Studios during the filming of The Lord of the Rings (2001), which was complex and much more restrictive, has now been obsolete
When filming the first film, Serkis wore clothes without tracking points, and the purpose was to facilitate the replacement of people with CG in the later stages. In the last two Episodes of The Lord of the Rings, Serkis finally put on a motion capture suit with track points. However, due to the fact that there are still a lot of obvious differences between the images of Serkis himself and Gollum, the general performance of Serkis is more of a skeleton reference, and most of the details need to be adjusted by the later personnel. In terms of facial expression capture, Serkis performed nearly 100 tracking points on the face, using close-up shooting material, and then handed over to the animators to deal with, like body motion capture, the effect of the expression needs to rely more on post-processing.
With the success of The Lord of the Rings, Weta Studios concocted the CG images of King Kong and Caesar from the Rise of the Apes series, and the technical principles behind them are the same, and the action-hunting actor behind these characters is this Andy Serkis.
Caesar, a CG character in Rise of the Apes
The movie's special effects big brother Industrial Light and Magic used similar techniques to create the CG image of Captain Octopus in Pirates of the Caribbean 2 (2006).
Captain Octopus in Pirates of the Caribbean 2
3, Aki - CG real people encountered "weird valley"
The flowing hair of the heroine Aki in Final Fantasy: The SoulSights (2001) was first realized
In 2001, there was also the first film to use motion capture technology throughout, and all the characters in the film used motion capture technology, which was Final Fantasy: The Soul Deep. The heroine of the film, Aki, has a soft and delicate hair fluttering in the wind. Never before had such a clear CG hair on the movie screen. In fact, the heroine Aki's hair is no longer made of dough, but first creates some threads to define the direction of the hair, and through these lines to generate strands of hair one by one.
Before Final Fantasy, it was extremely difficult to make hair with a computer and achieve its movement. This "Final Fantasy" should be said to be the first time in the CG world to achieve hair swing, rotation and dynamic light processing. In order to be realistic, the creators rendered a total of 60,000 strands of hair, which is already half the number of real hairs (ordinary people generally have about 100,000 hairs). The production cost was also quite staggering at the time, with a one-second rendering of thirty-six hours. It is precisely because hair is so difficult to create, so in the film, only the heroine Aki has long flowing hair, and the others are either short hair or bald heads, or wear hats, in order to "raise the short" and "avoid the long".
Motion capture technology in Final Fantasy
Final Fantasy was followed by films that used similar techniques, polar express (2004), Beowulf (2007), Ode to Christmas (2009), Mars Saver (2011), and The Adventures of Tintin (2011) by Spielberg, all of which were less than ideal at the box office. If you want to understand the reasons for the failure of these films, you can't help but talk about the "weird valley" theory.
The "Uncanny Valley" hypothesis was first proposed by Japanese roboticist Masahiro Mori in the 1970s. He believes that in the process of continuous personification of artificial humanoid objects such as robots, there is a possibility of a sudden decay in the process of increasing human emotions for them. That is to say, people's preference for robots is not directly related to the realism of robots, but when the emotion peaks, it will suddenly produce opposite feelings such as disgust, fear and even hatred. The chart below illustrates the gist of the "weird valley" hypothesis. In areas close to the intersection of the horizontal and vertical axes, people's liking for the humanoid body will increase with its realism; however, where the fidelity of the humanoid object is close to that of a corpse or zombie, people's favorability will drop sharply. This hypothesis is not difficult to understand. Moreover, not everyone agrees with this hypothesis.
Illustration of the "Weird Valley" hypothesis
Personally, I fully subscribe to the "Weird Valley" theory. From Final Fantasy to Polar Express, Beowulf, and Ode to Christmas, to The Adventures of Tintin, none of these films are comfortable.
The banshee played by Beowulf and Angelina Jolie in "Beowulf", both CG characters lack the texture of real people, making people feel strange and difficult to establish a sense of intimacy and identity
4, Benjamin Barton - CG character to achieve reverse growth
The protagonist in The Wonders of Benjamin Button has been growing backwards all his life
Facial expression capture technology is part of motion capture technology, which requires the performance of the performer's facial expression data to be recognized and processed by motion capture equipment, and then the digital information of the facial expression is bundled with the 3D model, so that the three-dimensional character produces the same realistic expression as the real shot.
The Wonders of Benjamin Button is the pinnacle of facial expression capture technology. The film is equivalent to a set of facial expression capture system tailored specifically for actor Brad Pitt, establishing an expression database containing more than 70 basic expressions, through which the combination of more than 70 basic expressions can produce almost any expression. Relying on this expression database can capture Brad Pitt's facial expression without a record point, capture and identify the expression of his real performance, match the expression in the database, reposition it to the model, and finally synthesize it with the body of the stand-in, which achieves the effect of "changing his head" and creating different images of Benjamin Button at all ages.
Facial expression capture technology plus "head swapping" created the CG character Benjamin Barton
Captain America 3 (2016) features young Iron Man Tony Stark. The special effects in this part are from the company Lola VFX. Before that, it was the company that used this rejuvenation technique in the movies Benjamin Button's Wonders and Ant-Man (2015). The basic principle of this technique is similar to modifying photos in Photoshop, but the specific details are much more complex. The young Tony Stark image in Captain America 3 lasted about 4,000 frames, and the special effects team had to take into account the change in every detail of the face, including the change of bones, from fifty-year-old actor Robert Downey Jr. to his twenties.
Iron Man Tony and CG Young In Captain America 3
After this technology matures, the characters in the film can become arbitrarily younger. In the Marvel movie "Captain Marvel," which will be released next month, we'll see the young S.H.I.E.L.D. director.
5. Paul Walker, Governor Tarkin, Princess Leia – the CG real person who "came back from the dead"
Paul Walker in Fast and Furious 7
The film "Fast and Furious 7" uses "head changing" to successfully achieve the rebirth of actor Paul Walker. Although the "head-changing technique" in this film is not as complex in creation as "The Wonders of Benjamin Button", the requirements for realism are higher. After all, The Wonders of Benjamin Barton is about creating an artistic image that doesn't exist, while Fast and Furious 7 is about "resurrecting" an established image. Judging from the final film, it is not at all clear which shot is the real Paul Walker and which shot is the post-synthesis.
The production process of the film is roughly as follows: 1. For field shooting, the film used four actors who were similar in height to Paul Walker to capture the main shooting data; 2, Paul's brothers, Cody Walker and Caleb Walker, used for close-up shots; 3, the actions of Paul Walker's stand-in were used as the base layer; 4. Matched Paul Walker's CGI face to the movements of these materials; 5. In order to create seamless output, Weta combined various CG rendering channels with the field shooting.
Above: Cody Walker as a "face replacement" (top), bottom: Cody (left) vs Paul (right).
Also using facial expression capture technology as the backbone of resurrection is the "Star Wars Story: Rogue One" released in 2017. The film's resurrection of actor Peter Cushing completes the most complex and expensive CG character remodeling ever made. Industrial Light & Magic took over the burden of this special effects task, and also with the help of the idea of the face database in "The Strange Story of Benjamin Barton", it established a facial expression library by extracting the facial features in the actor's past image data for making 3D models. Unlike Fast & Furious 7, the role of Governor Tarkin in this film is very heavy, and there are many close-ups of the actor's faces, so the requirements for the performance of the stuntman and the 3D model are more and more detailed. The crew specially invited the British actor Guy Henry, who was very close to Peter, as a stand-in, and he repeatedly observed the role of Governor Tarkin, completely imitating Peter's expression and performance style in front of the blue screen. When he puts on the motion capture system, the expressions and movements can be directly matched to the 3D models in the 3D production software. Then, simply cut out the avatar's head image and replace it with a 3D model of Peter's head to achieve the final effect.
Above: Stuntman wearing a motion capture system; Bottom: CG version of Governor Tarkin compared to the live-action Governor Tarkin
For films like the Fast and Furious and Star Wars films, a series of films may span years or even decades throughout an audience's life, accumulating countless iron fans. In the hearts of iron fans, the psychological impact of the change of an important role in the original play is very large, and this kind of moment to resurrect the deceased actor on the screen has become a heart-warming act in the hearts of the audience. At the end of the "Rogue One" film, the same screen resurrection technique is used to achieve the "resurrection" of the young Princess Leia, who is exactly the same as in Star Wars (1977) 40 years ago. This scene has made countless Star Wars fans cry and shock, for them, the spiritual shock and emotional comfort brought by the screen resurrection technique have far exceeded the need for plot promotion.
Princess Leia, who was "resurrected" in Star Wars Story: Rogue One
6, Alita - CG real people new heights
Alita: Battle Angel movie poster
Finally speaking of Alita. What is the technical difficulty of this movie?
First, the story of the film takes place in the real world, not like the alien planet or virtual world in "Avatar", Alita must achieve the same sense of reality as the real person;
Second, Alita is the protagonist of the film's story, not a dragon set character created by "head replacement" or "rejuvenation";
Third, Alita needs to get rid of the limitations of cameras and motion capture equipment and complete various difficult fighting actions.
This role can be seen as one of the highest level challenges launched by the world's top special effects companies to "Weird Valley". How much technical power was used to create this CG character? The entire film uses 30,000 computers, 800 special effects crews work for two years at the same time, and the overall rendering time adds up to 432 million hours. To make Alita look like a living character, the special effects team made her 132,000 hairs, 2,000 eyebrows, 480 eyelashes, and 8.3 million polygons in the "iris" alone...
In order to create the character of Alita, Cameron upgraded the motion capture technology and facial expression capture technology again, using a new "performance capture" technology. With this technology, changes in facial expressions and body movements can be recorded simultaneously, thus transcending the limitations of previously created CG characters. It is this technology that allows Alita to perform a variety of difficult moves that exceed human limits, creating a variety of exciting chase and fight scenes.
The "performance capture" technology upgraded to create Alita
Alita needs to complete a variety of combat moves that exceed the limits of the human body
After watching the movie, I will have different feelings about the role of Alita