Game Optimization Tricks (part 3) - LODs

When you're looking at what's on screen from a video game, you (as a human) tend to naturally utilize certain concepts like perspective to provide context for things like how far away things are from each other. Things farther away are represented as smaller on the screen, while things that are closer are larger. Naturally, this means that you can see things that are closer more clearly, and have a harder time seeing things that are far, far away. This is the founding principle behind the optimization concept called "Level of Detail", or "LOD".

What kind of designer should I be? The Cinematic Designer (part 3 of many)

It has been a while since I've done one of these, so I wanted to return to it with the new year. There are a lot of kinds of designers out there, and one of the most thankless but most important roles is that of the cinematic designer. Others might call them a camera designer, but this job entails the visualizations of how things play out on screen. It's a younger role in the industry than some others like level design, because the transition from 2 dimensions to 3 dimensions brought about a new concept - what the camera looks at. Previously, 2D games were much simpler, mostly with a mostly fixed camera position and sprite-based graphics. Once 3D became a thing, suddenly an entirely new world opened up, and game designers could start incorporating techniques from film as additional tools to help craft a more directed user experience. This meant that games, especially in 3D, have moved towards a more cinematic experience, which creates a need for designers who specialize in this sort of visualization.

Cinematic designers often work very closely with animators to help craft specific scenes. The animator works on the characters involved, and often has ideas about how the scene should be framed, but it is generally up to the cinematic designer how to evoke a specific theme or feeling via the camera and how the scene is viewed, and there are a myriad ways of viewing the scene, each with its own idiosyncrasies. Cinematic designers are a bit of an unsung hero in the industry. Camera systems are a lot like UI - you never notice a good camera system, despite the fact that it lets you see what you want when you want it, but players everywhere know the pain of wrestling with a bad camera system.

A large part of the job is just dealing with the applied constraints - you cannot allow the camera to penetrate world geometry, clip through characters, or see outside of the world. That will result in certification failure. There are other softer constraints as well - camera movement that is too rapid or too jarring can be disorienting for the viewer. Then there are technical constraints as well - can this scene occur anywhere? What if the characters are on a slope? What if they are near a wall? Where will the camera be relative to them? Will it cause issues? How do you handle it if you need to play a scene anywhere, but it requires a certain amount of space around the characters for the camera? Solving these problems efficiently is definitely not trivial.

Beyond that, you also have to consider the sort of tools and effects you have at your disposal as well. Camera attributes also have a tremendous effect on what's being viewed. You can make objects look far away or nearby simply by changing the camera focal length. You can force the viewer to focus on specific items in the foreground or background by changing the depth of field. You can add camera effects, like motion blur or camera shake in order to try to convey specific effects to the player as well. These are all tools that a cinematic designer can use to craft an experience meant to evoke specific themes and feelings.

This is an example of a "Dolly Zoom" shot. Not every camera shot is as noticeably weird as this, but definitely it shows what sort of effect that just the camera can have on a scene. The way this shot is made is by moving the camera towards or away from the subjects while adjusting the zoom on the lens at the same time. It creates a sense of vertigo or "strangeness" in the viewer, because things that they expect to be one place are moving, while other reference objects in the scene are not. Think about when you would want to use a shot like this in a game, what sort of situations might you want to try to disorient the player? 

Perhaps the biggest difference between traditional film making and cinematic design is the amount of time you have to work with. For gameplay, you often have to deal with something as quick as a fraction of a second. You need to tell a simple, yet believable story to the player within the space of 30, 60, 90 frames of animation and all you have is the sound effect, camera, and animation to work with. How do you do something like that? How do you make it work? And on top of that, the animations and camera you're using are often repeated many times over the course of the gameplay. How do you keep it from getting boring?

Take a look at these takedown sequences from Splinter Cell: Blacklist. If you notice, they're all occuring over the span of a few seconds, never more than 5 and most never going longer than 2-3. As an exercise, I want you to think about why some of them have camera movement and why some of them don't. What does the camera movement add? What sort of feelings does it evoke? More importantly, what is it the viewer gets to see while it is going on? How much is focused on what the player sees, and how much is focused on what other dangers are within visible range as well?

Does this sound interesting to you? Would you like to be a cinematic designer?

For students interested in this sort of work, I especially suggest studying filmmaking techniques. Know your tools, know your shots, do the homework - a lot of these are classic tools used in filmmaking, and adopted to work in the medium of video games. Also, try to take courses in animation - both traditional and computer-assisted. You need to know how things move over time in order to best display it. 

For those who aren't in school but interested in this sort of work anyway, I suggest the "fun" homework. Go back and watch classic action movies and note exactly what the camera is doing during their action sequences. Try watching just the action sequence with the audio turned completely off, and only watch the camera. Scrub through it frame by frame if you can and try to isolate the sort of techniques they are using as well as what effects those techniques have on the viewer. They aren't big on their own, but they will definitely add up, and I guarantee the viewer's subconscious brain will see and react to it, even if the viewer doesn't consciously recognize it.

The purpose of controlling the camera is to help guide what the player sees. You want and even expect that you need to relinquish control to the player some or most of the time, but even then it needs to be handled with care, because you want to help the player see what he or she wants to see. You want to make it easy for the player, to reduce the frustration as much as possible, and still be able to help tell a cohesive story, even if that story is only 2 seconds long, and consists of "James Bond takes a thug down". If you find it an interesting challenge to try to convey to the viewer that James Bond is a badass, that he is strong, powerful, capable, and stealthy, that he is a real terror in a fight, and do it all within 3 seconds of animation and camera work, this is the job for you.

So you want to be a game designer (part 3 of many)

One of the most common mistakes I see novice designers make is that they don't focus enough on making the game mechanics and the game narrative mesh well together. It was something that was mostly forgivable twenty or thirty years ago mostly due to technical limitations, but it severely hinders the flow and immersion of the game when it is done. This is the sort of thing that outside observers will typically remark on - pointing out how "video gamey" something is because it doesn't model certain expectations in some way.

Here's one example. At one point during development of the more recent Tomb Raider games, the designers needed to figure out how Lara would regain health that she had lost to combat or falling or traps or whatever. They debated whether she should find glowing health pickups (which would detract from the pseudo-realistic presentation and aesthetic), or finding medical kits (which would cause the players to question why there were medical kits buried in ancient tombs). At one point, they had actually decided that Lara would perform a gymnastics floor routine in place to heal from her injuries. Thankfully, that was also left on the cutting room floor. Eventually, they decided on the old standby of regenerating health instead. It still had the problem of hurting believability - why would Lara recover from bullet wounds simply by standing around? Nonetheless, it shipped that way because it was judged to be the solution that introduced the least problems compared to the problem it solved.

The important thing is to reinforce what the game's narrative is about through the rules. It doesn't have to be about the big story, but the rules and mechanics of the game should reinforce what is happening. The human brain sees these sorts of reinforcements and makes that 'a ha!' connection, even if it is subconscious. As a designer, you always want to leverage the existing knowledge in your player's brain to your advantage, because it helps create believability and immersion by filling in the little mental gaps with their own knowledge of the topic.

The best place to look at this are games where there isn't much more than the rules. No fancy graphics, no mini games, using only the rules and mechanics to show what the story is. This is harkens back to what I wrote about in [part 2 of “So You Want to be a Game Designer”] - play a lot of games. And I don't just mean video games either, but board games and card games alike. Tabletop games especially are all about using mechanics to convey the narrative. Like this card, for example - it makes perfect sense. The demon grants three great benefits, but your doom draws closer with each benefit you take. This card tells its own story through its mechanics, and it feels exactly like an ordeal you would expect a wizard who has made a demonic bargain to go through. 

As another example, let’s look at the game Robo Rally. It's a fairly simple game - you have a map and pieces. Each player controls a robot and they are trying to reach a certain location. So how does it work? Each player puts down a series of instructions for his or her robot to follow - sequentially. But other things can happen during the game. The robots can run into each other, or get knocked out of position. They can fall into hazards, or other things can happen. But their players can't instantly course-correct, because the robots are simple creatures that only run their commands sequentially. So if something happens to the robot that would put it out of position, it will continue executing its orders to the best of its ability, even if that means trying to run into a wall, or going the wrong way.

This makes a ton of intuitive sense to anyone who's tried to tell a robot what to do. If you replaced the robots with mice, soldiers, unicorns, or some other creature, it wouldn't make as much intuitive sense to the player. Why would a vampire run into a wall? Constructing a narrative of robots going haywire and bumping into each other works hand-in-hand with the mechanics of the game. It makes it much more intuitive because it relies on the understanding the player brings regarding robots.

This concept also works against you when mechanics and rules fight against the intuition, which breaks immersion quite easily. If you've played Final Fantasy XII, you may have felt it yourself. Over the course of the game, the characters level up and earn experience points which can be spent acquiring "licenses" to use items and equipment. But that breaks all sorts of intuitive sense when your character can't even wear something like a leather hat, or a cotton shirt without spending license points to obtain the license for it.

Blending the game's narrative and mechanics is one of the things that separates the armchair designer from a real designer. It's incredibly easy to be just read about it and be critical about them, but much more difficult coming up with narratively-consistent elegant solutions of your own. How would you go about creating a game mechanic that properly expresses a specific and abstract concept, like speed or heat? How do you make a dragon seem large and dangerous if your game doesn't have a concept of combat?

As a homework assignment, you should think about trying to come up with game mechanics that can be used in a limited environment (think board or card game) to express narrative concepts like:

Once you come up with something, try playing it. See if you can make it work. Which mechanics worked? Which didn’t? Can they be improved? These are all extremely valuable skills for any game designer to have, no matter what sort of game you're working on.