An action scene's quality is determined by a variety of factors: coherence, both spatial and narrative; investment in the characters; and choreography. Imagine, if you will, that a good action scene is a successful Rube Goldberg machine. All of the elements work together, cascading one after the other, escalating in stakes in order to keep up suspense. An action scene requires coherence in order for the audience to follow, visually and narratively. The spatial and narrative relations between the pieces of the Rube Goldberg machine must be clear at all times. An incoherent action scene is full of sound and fury, signifying nothing. If the narrative stakes are not clear, then the action scene is an empty vessel, telling no story beyond a series of disconnected images that concatenate to nothing. If the spatial relations are not clear, the audience cannot invest in the narrative as they cannot follow the narrative onscreen.
Another element of the action scene's quality is the conditions of production and overall aesthetic. Not all action scenes are the same, regardless of quality, and thus do not all look and feel the same. The visceral confusion and chaos of Paul Greengrass's The Bourne Ultimatum is not the same as the crisp and clean of Adam Wingard's The Guest or Chad Stahelski and David Leitch's John Wick. The rapid fire cutting of Edgar Wright's The World's End differs from the same quick cuts of Gareth Evans' The Raid: Redemption. Notice that the conditions of production are vastly different: production budgets widely divergent, countries and thus modes of production varied. Different cinematographers, production designers, stunt crews, editors, etc all change the feel of the action scene. Their overall aesthetics are different. This is a good thing. Homogeneous aesthetics can only have diminishing returns. The aesthetics of Furious 7 and The Avengers: Age of Ultron (hereafter F7 and AOU) are superficially different (colour palette, lenses, depth of field, etc) yet feel remarkably similar. However, F7's action scenes manage to be successfully coherent in both spatial relations and narrative whereas AOU suffers from fatal incoherence.
Yes, it's time for my belated #hottake on AOU in which I bemoan the state of affairs known as the mainstream Hollywood blockbuster, a topic that I return to again and again. However, this time, I have some positive things to say about the blockbusters themselves -- though I retain my ire for the production paradigm.
So far in this summer movie season, I have seen F7, AOU, Mad Max: Fury Road, Entourage, Ex Machina and a couple smaller movies. Mad Max: Fury Road is not only the best action film of the year for me, it's also probably the best film of the year. It's visceral, well-made, lean, and feminist as fuck. MMFR is the feminist action masterpiece I've been waiting for my whole life. It was like coming home. The release of MMFR has sparked endless thinkpieces about the muted response that greeted AOU and the rapturous praise garnered by MMFR. Many of them focus on the supposed exhaustion over the Marvel Cinematic Universe and its imitators. Writers and critics pontificate on the logic of anticipation and the diminishing returns from such a model. Some thinkers have opined the qualitative difference between the two films is one's reliance on practical effects over CGI heavy scenes.
Over at Variety, Brian Lowry writes that:
The pattern has become predictable. “Iron Man,” a terrific movie overall — particularly in capturing the origin story — degenerated into a mundane brawl between two armor-clad characters. Ditto the “Hulk” reboot with Edward Norton, which culminated with the title character’s ho-hum showdown with another green behemoth, the Abomination.Notice that this post has already used "diminishing returns" three times, including Lowry's use. The common complaint about these blockbuster movies is that the climactic scenes are simply unintelligible, turgid, or boring action scenes that unfold endlessly. The stakes aren't pronounced as these blobs of pixels fight other blobs of pixels with blobs of pixels between them. Lowry continues to say that AOU "becomes drearily repetitive as the heroes mow down another CGI horde, this time consisting of artificially intelligent robots[,]" the minions of Ultron, the film's titular antagonist, voiced by James Spader, but given no coherent motivation or compelling characterization by the film's script.
One can argue, in fact, that the much-maligned second “Star Wars” trilogy sacrificed an element of its humanity in George Lucas’ embrace of a wholly digital filmmaking approach. At a certain point, watching droid armies being whacked to pieces begins to yield diminishing returns.
Put more simply, just because CGI wizardry allows you to do something, whether hoisting an entire city into the air or leveling skyscrapers willy-nilly, doesn’t always mean you should. Because while the box office figures might suggest otherwise, there is an audience out there that’s weary of these movies precisely because of the hollow quality to the inevitable final 30 minutes of unrelenting mayhem.
Ultron's plan, if it can be said to be a plan, is to "protect" humanity by wiping it off the face of the planet using a levitated city that he will drop using a countdown mechanism or a big red button, I can't remember but it doesn't matter because nothing in these films really matters in terms of narrative stakes. I'm reminded of David Bordwell's fantastic commentary on the structure of action movies. He and his critical partner Kristin Thompson outline five guiding principles of storytelling (as they dub the category):
- Goal orientation
- The double plotline
- Discrete part-structure
- Planting causes for future effects
In AOU, the characters have multiple goals: Tony Stark is interested in retirement by automating a system of protection, Steve Rogers is interested in eradicating the subversive and corrupting elements of HYDRA and other bad guys, Natasha Romanoff is focused on achieving a semblance of normalcy by a romantic relationship with Bruce Banner, Wanda and Piotr's goal is revenge against The Avengers, specifically Tony Stark for his complicity in their parents' death, and Ultron's goal is world destruction. Thor doesn't really have any motivation because the script has to sideline somebody when they are so many characters. The goals become more articulated and narrow as the film's plot progresses: the primary goal is to thwart Ultron.
The film features a complex but still discrete double plotline: 1) The Avengers fight Ultron and 2) the characters struggle with the divide between their personal lives and their responsibility to abstract concepts such as justice or revenge. For example, Natasha and Bruce would like to fall in love or pursue a romantic relationship but they each find themselves morally repellent (Bruce accidentally kills people, Natasha was programmed to kill).
The discrete part-structure dictates that, David writes, "[t]he action revolves around goals: defining them, modifying them, and achieving or not achieving them. Hollywood films map the process onto several parts, each running 25–35 minutes." In the first discrete part, The Avengers invade a HYDRA base with the goal of recovering Loki's staff. In the second discrete part, Tony and Bruce engineer Ultron with the goal of retirement. And so on and so forth.
A classic example of planting causes for future effects, or a Chekhov's gun, involves the famous party scene in which each Avenger attempts to lift Thor's hammer. Only the worthy are able to lift the hammer. This is both a plot and a thematic seed that is reaped when The Vision demonstrates that he is worthy by lifting the hammer, thus proving that he can be trusted (a plot point) and that his creation is worth the cost (a thematic point).
Finally, the climax of the film features a deadline: The Avengers must stop the levitating city before it reaches a critical altitude or less the city's impact with the planet will cause a mass extinction event (an idea thoroughly debunked here). Ultron will touch the big red button at the critical point, the terminal altitude, if you will, and must be stopped.
In the same article, Bordwell cites Thompson by contending that, "[i]n features running around two hours, we typically find a four-part structure: Setup, Complicating Action, Development, and Climax. Usually there’s a brief epilogue tacked on. Filmmakers working in the three-act paradigm in effect split the second act into two stretches around a midpoint." The Setup in AOU involves the invasion at Hydra and the creation of Ultron. The Complicating Action is, of course, Ultron going evil and marshalling Wanda and Piotr. The Development involves The Avengers trying to get ahold of Ultron's new body and succeeding, resulting in the creation of The Vision. The Climax is the long CGI fight on the levitating city. The Epilogue introduces the new team of Avengers. This broad outline is of course reductive, but the film does quite neatly fit into these broad categories. However, as with many bloated blockbusters, AOU is 140 minutes and thus each discrete part eats up a little bit more time.
None of these observations I've made using Bordwell's basic formal methodology are qualitative judgements. Stories that use appropriate and familiar structures are not to be dismissed on the basis of their use. Rather, it's like any genre exercise: if it's done well, then any formulaic aspect can be forgiven or minimized in terms of judgement.
I really don't think AOU does any of this particular well. In fact, it feels all rather rote, and this is because there are little narrative stakes thanks to the logic of the escalating blockbuster. Allow me to quote myself on the logic of escalation:
The films themselves present themselves as both the object of desire and the cause of the desire for the desired object. Each item in the series of advertisements (another endless wave of production) simply teases the inevitable tease that is the film itself.The world was at stake in the first Avengers film and again, the world is at stake in the second Avengers film. Each film must both reset the status quo (The Avengers themselves continue to fight another day) and set up the next antagonist (Thanos continues his mind numbingly dumb plan to assemble the Infinity Gems or Stones). The films are, in essence, a holding pattern with the illusion of increased narrative stakes. They bide time while increasing the audience anticipation for the next film, the next "world in peril" problem that we know will be thwarted. The problems of the previous films are diminished as the problems of the future unreleased films tease their importance. The audience strains to take seriously the threat of Ultron when around the corner (three years in the future), Thanos bides his time with his nonsensical plan.
This constant anticipation has, of course, an unintended side effect within the overarching structure of the production of blockbusters. Each film, independent of intellectual property or studio, must somehow compete with the very aesthetic. That is to say, that each subsequent item in the endless series must somehow replace and better the previous item. The stakes in films get higher with each new release.
It doesn't help that Ultron is a poorly drawn villain and a poorly established threat. If this is a sentient AI that is endlessly self-replicating, it strains credibility that he would put all of his focus on a single half-organic, half-cybernetic body. It also strains credibility that his army of drones would be so... weak and tiny. His AI has the ability to stretch out infinitely as he can colonize computing power for increased numerical superiority. The Avengers themselves should have been thoroughly overwhelmed by sheer numbers. Yet, Ultron puts all of his eggs in the Vision's casket.
Lowry writes that, "there is an audience out there that’s weary of these movies precisely because of the hollow quality to the inevitable final 30 minutes of unrelenting mayhem" and I put myself and many others firmly in that camp. The hollowness of Ultron himself echoes out to the entire film itself. It's as by the numbers, formally speaking, as the worst of its ilk.
A great example of the formal mishandling of this film, then. Unfortunately, I cannot find of a screenshot of the film (and hopefully when the film is officially released for home use, I can remember to come back and delete this part and include the shot) so you'll have to bear with my description.
Natasha Romanoff has been captured by Ultron and is being held in his underground lair. Seemingly hundreds of his drones scramble around building the thrusters for the levitating city trick. Ultron speaks through the prison bars to Natasha in an attempt to scare her. This Ultron's head is crushed by the hand of a bigger Ultron. This scene has a few purposes: to demonstrate the increased stakes of a superior Ultron body; to indicate Ultron's ruthless nature (he would destroy his own body); and to be a "cool" moment, memorable to the audience.
This scene fails on every level it attempts. Firstly, the destruction of his own body is wasteful and nonsensical. The smaller body could still function as a drone, as an even better version than the weak drones Ultron uses. The increased stakes aren't articulated very well as the crew shoots this scene in a nonsensical manner. Ultron is a motion capture CGI creation, but he shares his screen with a real life background. The crew chose a particular lens that puts the background in least amount of focus. Thus, the relation between the figure and the background become distant. The distance of the background creates ambiguity in the size of the figure; with no point of reference, the figure could be any size. In order words, the practical increase of size that Ultron is supposedly demonstrating is nullified by the ambiguous spatial relation between the figure and the background. He's supposed to be bigger but because we have no frame of reference to know how much bigger, the scene fails.
While Whedon might be great with dialogue ("please be a secret door. Yay!") and, most of the time, character, he is not a visual director. The medium of film uses a wide range of artistic tools in order to convey important information, whether that info be affective, metaphorical, plotting, or data. An excellent filmmaker uses all aspects of the medium: editing, sound, composition, dialogue, etc, etc, etc. Whedon is not a filmmaker that uses all of these elements together in a beautiful or affecting manner. Rather, his sense of the visual is muted or even blunted. The failure of the scene of the bigger Ultron is repeated throughout the film.
In one discrete part of the film, Ultron is attempting to abscond with the creche that holds the nascent Vision. The Avengers' goal in this scene is to capture the casket. In order to accomplish this goal, Captain America plans to distract the main body of Ultron through hand-to-hand combat. The scene depicts Cap and Ultron fighting on a train. From a shot-by-shot basis, it's not always entirely clear where Cap is in relation to Ultron, or even where they are on the train itself. The audience is left unmoored from the spatial signifying chain through inept intercutting between Natasha on a motorcycle fighting drones and Cap fighting Ultron on the train. Of course, Ultron escapes (this is the Development part of the film) and Cap is left to stop the barrelling train... which inexplicably comes to a dead-end with a station in between. It's pure nonsense.
It goes on and on as bundles of pixels fly across the screen against other bundles of pixels. None of the specific shots are memorable as composition is not regarded or even considered it seems. Many shots are dizzyingly short, too short for the human eye to consider and evaluate the studium and the punctum. The action scenes fail because the stakes are either nonexistent or unclear. Thanks to the logic of blockbusters, the threat of Ultron can never be established fully. The film is a holding pattern and thus each action beat, each note of each bar, is essentially an empty moment, a rest camouflaged as a note. Each punch is meaningless.
Again, the formulaic structure of the individual film is not the problem. The rote motions of the plot could be exciting or successful if the full spectrum of filmic elements were mobilized. Recall my use of the Rube Goldberg machine. Each element cascades into the next, each moving part affects the next, requires the previous. It's the difference between the inferior "and then" between moments and the superior "because" between moments. While AOU might use the storytelling principles that other films use, it fails as a whole because nothing matters in the context of the Marvel Cinematic Universe and nothing matters within the individual film itself. "Cool" moments are strung together in a discrete linear fashion, where each "cool" moment sits unmotivated to affect the next. It's just so boring. If nothing matters, then the audience is given nothing to care about.
I cited F7 up at the beginning of this post. It's, in every way, a better movie than AOU. Not only because James Wan knows how to use more tools from the filmic toolbox, but because the stakes are always crisp and clear. The best scene in the film is the attack on the convoy. In the film, during the Complicating Action stage of the structure (this film slavishly follows the structure just as much as AOU), the ersatz crew is tasked with liberating a hacker from a heavily guarded and heavily armed convoy of trucks. The goal is clearly articulated by exposition. The trucks are moving through a mountain pass that has no intersections and the crew has no way to intercept other than to release cars from the hold of a plane in flight and land them on the road (impossible but always clear). Then, the cars must quickly drive up the road, intercept the convoy, liberate the hacker, all before the convoy reaches a critical spot in the drive and thus the crew would lose that important and solitary window of opportunity.
During the sequence, Wan cuts between five muscle cars on the road, the four or five trucks, and the single personnel carrier in which their quarry is confined. As each part of the plan progresses, the fortunes of the heroes is overturned. Each moment features a complication, each time increasing the stakes as the goal becomes increasingly more difficult to accomplish. One of the muscle cars fires a harpoon into the back of the personnel carrier. The back of the carrier is ripped off. Paul Walker's character, Brian, must jump from his car into the carrier. However, the carrier is armed and fires machine guns at the crew, ie a development and then a complication. Once inside the carrier, Brian's task is not finished as he must release the hacker from the cage using the computer. A development. Then, a complication in the form of Tony Jaa's character Kiet. He seeks to stop Brian; they fight.
Each moment increases the stakes as the complications increased in their complexity. The incredible scene culminates in the carrier on its side going over the edge of a cliff, Brian running across the side of the carrier and jumping off at the last moment onto Letty's car. However, Letty's car did not come from nowhere and raced through its own series of complications including armed trucks with blazing guns.
Wan deftly intercuts between each character's complications and their efforts to overcome the obstacle. Each character is given something clear to do and has a clear throughline. Each complication increases the stakes for each character in a clear manner (such as for example, one of the parachutes not deploying correctly leaving one car to catch up).
It's one of the finest action scenes I've seen in a long time. Consider then, as an alternative, AOU's climactic action scene.
In the city centre, Ultron has set up his big red button. The Avengers invade the city with the goal of stopping Ultron from dropping the city. So far, no problem: clear goals, clear geography: get to the city's centre. As the fight progresses between Ultron's drones and the team, complications occur. One of these complications is that Cap is determined to evacuate all of the civilians no matter what. Another complication is that Wanda is facing a bit of cowardice. None of these complications build in terms of a Rube Goldberg machine. Wanda's cowardice doesn't get overturned into another complication with increased stakes, either narrative or thematic.
At one point, The Vision and Ultron's main body get into a fist fight with Thor sort of on the sidelines. Once the scene comes to its natural end (The Vision punching Ultron away from the machine, giving The Avengers space to control the space around the machine), The Vision is sidelined in terms of screen time. We do not know what happens to The Vision. He and Thor are so overpowered that it makes narrative sense to keep them offscreen; otherwise they would dominate the fight and there would be no tension. Thus, Whedon keeps him offscreen. It makes sense logistically but it's quite disappointing narratively speaking.
The same happens to Hulk. He gets punched and thrown across the city. We don't follow up with him for what seems like ten minutes. What is he doing in that ten to fifteen minutes of diegetic time? It of course doesn't matter. The audience is always unsure of where the Hulk has landed -- this is especially unclear when Black Widow, while fighting drones in the streets, just comes across him. How far was he thrown? How far is the city centre from his point of impact? Who knows? What this demonstrates is that Whedon is unable to juggle all of the elements of the action scene coherently. He is unable to provide clear goals from one scene to the next and he is unable to string those moments together in a meaningful or suspenseful way.
The climax culminates in The Avengers surrounding the machine and holding at bay the endless drones. The camera circles around the team showcasing each one hitting the drones. The credits of the film depicts this moment as a statue. This begs the question: how dynamic can a cinematic action sequence be if you can easily depict it as a static, unmoving, never changing statue?
Action scenes depend on spatial awareness, narrative stakes, and clarity. The Avengers: Age of Ultron is not great at any of these things. It's not even good at setting up the next crisis. In a scene already infamous, during the Developing part of the film, Thor takes off from the team, picks up that scientist from the Thor movies, strips his clothes and takes a bath in a cavern. This submerging provides him with a vision of the future or something blah blah blah. It's stupendously unclear and the scene's purpose is twofold: provide Thor with the necessary information for the creation of The Vision and to set up the vast and terrible threat of Thanos. It is amazingly unclear how the vision (which is of the Infinity Stones) is able to give Thor the information on how to bring The Vision to life. Which he didn't even know about. Because he took off. Ugh.
In other words, the movie is far too busy with incident and yet paradoxically empty with incident. Nothing happens but lots of things happen. It's 140 minutes of discrete moments that do not concatenate to a meaningful experience. And it's not because it's formulaic or because it's a mainstream Hollywood blockbuster (because F7 and MMFR seem to be successful). It's the utter inept manner in which it goes through the motions. Other than Entourage, AOU has been easily the worst film I've seen all year.
I certainly didn't go into AOU expecting to hate it as much as I did. The first one was okay enough. Guardians of the Galaxy and Iron Man 3 and especially Captain America: The Winter Soldier were all decent enough. These movies are good examples of how to just simply be good at the bare minimum. AOU just doesn't. It's awful.