Tuesday, September 20, 2011

Law Of Diminishing Returns: Redux

This post is a revisiting of a previous post from this blog, and is also posted here.

Every so often, the hot topic of crunch/overtime/extra hours comes (read: anytime a jackass like me writes a post about it or the mainstream media decides to interpret what a jackass like me writes as a scathing commentary on the state of the industry). Sometimes, though, high profile people say something about overtime, and it incites a reaction, no matter what the intent.

It’s no surprise to anyone, in any industry, that overtime can creep into all facets of our lives. In the case of me being both a powerlifter and a video game developer, overtime has come in the form of overtraining and crunching, respectively.

Not this kind of crunch, guys

For powerlifters, overtraining generally involves training your body so hard that it effectively cannot recover- your basal body temperature drops, your central nervous system begins to shut down. For game developers, crunching is typically a point in a project that requires developers to work 60, 80, or even 100 hours a week to hit a deadline, pushing our mental capacity to a limit. In both cases, you become lethargic and your sleep suffers, furthering the lack of recovery. The belief for both overtraining and crunch is that that the extra time put in will result in higher quantity and quality of results.

Our ability to cope with such overtime hinges heavily on the amount of time we have to recover from engaging in it.This recovery, whether or active (by, say, deep tissue massage or engaging in an unrelated interest) or passive (by relaxing on the couch or playing video games), can only work if those needing it are given the proper amount of time and resources to ALLOW it to work.

All The King's Horses

Recently, I spent a large amount of time training for a national powerlifting competition. I trained for 5 months, competed in a local meet as a warm up, and then trained for another 2 months before the national meet. During the initial 5 months I worked very hard in the gym and took regular breaks from training so my body could recover, and I saw great gains in strength. At the local meet I set a personal best in the deadlift, and walked away happy.

After the local  meet, I took only a few days off, and then decided that I had recovered from the first meet and the stress it had put on my body.  Since I had limited time before the national competition, I spent the next 2 months working even harder than I had before. Not only did I not take enough time off after the local meet, I eliminated any down time from my training, effectively giving my body no time to recover. After several weeks of this new training plan, I actually felt weaker. I started skipping workouts and making excuses to myself to not push it too hard. It got to the point where I didn't even want to compete in the national meet anymore. I was burnt out, stressed out, and flat out tired. Even still, I set a goal to break my personal record for total weight. In fact, I planned on pushing myself to move more weight than I had ever done, even during training.

He probably overtrained something...

So what happened? I failed. I didn’t allow my body to recover from the incredible stress I was putting on it, and as a result when it mattered most I couldn’t compete at my full potential. I didn’t reach my goals, and the national meet was one of my worst in recent memory.

All The King's Men

Game devs are probably wondering what this meathead, lunkhead, Planet Fitness-reject's story has to do with them. Well, to paraphrase Matthew McConaughey in “A Time To Kill,” just imagine if what I described above was the progression from a press demo to a final build (or, maybe the progression from shipping a game to working on the next one).

Imagine that the whole team gets behind polishing the absolute hell out of the first hour of gameplay. We spend a good 6 months of time getting the first 3 missions right. The scripting, the code hooks, the character performances, everything. We show it to the press during the months leading up to launch, and all of our attention is focused on that one hour of gameplay shining like a star.

After all of that, the realization sets in that we have 3 months to finish the game. We have to get ALL of that hype into the rest of the missions, the gameplay, the general feel of the game. Maybe this mission here doesn’t have enough of what dude from Kotaku loved, and maybe those animations there don’t live up to the expectations of our publisher. So we start spending extra time revamping, changing, cramming in content. Features creep in. Hours start piling up. Some of us are thinking "I'm working extra hours because I believe in the project," so when we hit a wall at 10pm, we push through it.

We probably looked like MC Hammer on crack at times, too

Suddenly it's Thursday, 2 weeks before submission, and we've already put 50 hours in this week. We start to check in bad data. We break the build. We snap at co-workers. We're no longer being smart or creative about making the game- instead we're going through zombie-like motions to just get it done. We just want to ship this thing and move on to the next project.

And that passion we had? Starting to dwindle, if not gone already.

Sound familiar?

Putting It Back Together

The thing is, though, that it doesn't have to be like this. There can be good overtime. For example, I've overtrained my deadlift and seen incredible gains by pushing myself on heavy days and taking a month away from the next heavy session. I've worked 50-60 hour weeks, with 5 (and rarely 6) days a week for no longer than 2 or 3 weeks to get a deliverable out the door, and have produced work that was higher quality and more rewarding, with no negative effect on my health or marriage.

What made the good overtime better than the bad?

After the bad overtime, I was done. I thought I had recovered enough, so when I tried to (physically and mentally) get back into both activities, I couldn’t. I was done, and wanted out. I considered quitting both powerlifting and game development after the last bad overtime experience with each.

In the case of the good overtime experiences, I was able to take the proper amount of time off that both my body and mind needed to recover not just AFTER the overtime, but DURING it. The work I put during these smaller pushes was of higher quality, more rewarding in the end, and most importantly, kept me engaged in I was doing and looking forward to getting back to it at full tilt as soon as I could.

It can be argued that if we want to be successful, we have to push ourselves harder than the average in our fields. It doesn’t, however, have to have a negative affect on the things we are passionate about. We’ve all read the reports, seen the opinion pieces, heard about EA Spouses and Kaos’ “thousand yard stare.” I’ve read articles on how overtraining has blown out knees, biceps, backs, and worse. Everyone universally agrees that too much overtime is bad- Bad for your health, bad for relationships, bad for studio morale, bad bad bad.

Smell The Roses

Don't let life pass you by
Overtime exists and it’s not going away. I’m not suggesting that it does. I’m not going to rant about crunch time ruining lives. I’m not going to claim that my life has been horribly affected by working overtime or training too hard.

I am, however, going to say this- we all need to manage it better.

I don’t mean that we need to plan better (we know), or avoid feature/exercise creep (we try), or never put in overtime (we will). I mean that we as individuals need to manage how we represent ourselves while working overtime. We need to be conscious of the fact that people who are interested in what we do (powerlifting, game development, insert-your-interest-here) are going to look at us as an example. They’ll see us doing stupid things in the gym or working 100 hours a week, and see us wearing both of those things like honor badges. They’ll see us tweeting about how we’re “crunching to make the game better for you, the consumer!”, or read our Facebook post about how we just totally killed a training session and can’t walk right now- but hey, "no pain no gain!"

Those people will enter our fields and expect that to be the norm, the right way to do things, and they will never question those methods until they too are burnt out. And that’s a damn shame, because we can prevent it. We can teach these newcomers a different lesson- to not make the mistakes that we did. We need to encourage them to come into our industries and change them for the better.

When all is said and done, people will only remember the 4-million-on-day-one sellers, and not the people who worked hard and sacrificed to get the game to that point. We’ll only remember the monster numbers that a powerlifter put up at Worlds, but we’ll never see the training that was put in to achieve that. So let’s take back that part. Let’s do it smarter. Let’s follow the Law of Diminishing Returns.

Monday, September 5, 2011

Technically, It's Creative!

This blog was cross-posted here.

While waiting for a reply to my open letter from a few weeks ago (I still haven’t heard back from you, Andy!), I gave a talk at PAXDev regarding how to animate quickly for design, which contained both creative and technical approaches to game animation. After the talk, I received an email asking two things:

 “How do I get into video game animation? I've always found animation to be a very interesting discipline. The problem is that I'm not even sure where to begin training for such a thing.”

“Is animation more technical than artistic, or is it the other way around?”

The first question is something I am asked often, and am glad to help with. I responded to the inquiry with this:

“The first thing to do is learn the basics- read up on the 12 Principles of Animation (read The Illusion of Life- it's the animation holy bible). If you can draw, you could start doing the basic animation tests (bouncing ball, flour sack, etc.) that way w/o needing to buy Max or Maya (though both have 30 day trials!) I've seen far too many beginners jump right in with a walk cycle and not be able to take or understand critique on their work because they don't understand the principles and the vocabulary.

http://www.11secondclub.com/forum/ is a great place to share animation, as well, once you feel comfortable with it. Online schools like http://www.animationmentor.com/ and http://ianimate.net/ are good, but they are not cheap.

Also, watch movies! Watch old WB classics, watch Disney, Pixar, and study the animation. If you see something cool, rewind and step through frame by frame to see what they did. Deconstructing the masters is a great teaching tool.”

However, second part really made me think. Is animation more technical than artistic, or is it the other way around?

The easy answer to this very good question is “yes.” Game animation is both creative AND technical. Animation is creative by nature-we’re figuring out timing of our motions, developing poses that communicate emotion and intent to the player, and ultimately, we’re delivering the look and feel that design wants. In all, we are breathing life into our characters!

Nah, we're just painting pixels, really

However, it helps if you have some proficiency with understanding the technical side of things (though I’m not talking about rigging and other technical animation specialties - those have their own discipline separate from animation). In some engines, your timing in that walk cycle needs to match the timing for the rest of your walk cycles so it blends properly, or your footstep events need to be setup so time-scale blending to your run cycle works properly. In other engines, your poses need to be setup so that they makes sense against other poses they will blend to, unless you want to create countless 3 to 6 frame transition animations.

And you better know what state machines and blendtrees are. If you know how to build them, that’s a big advantage. Know a little scripting? That’s even better.

Oh hey, I can handle that!
This begs the question, however, of where exactly the line drawn is between the two, and how important it is as an animator to BE technically trained. Sure, you can create the best looking run cycle in the history of run cycles, but how does it work with animation layers in-game, how does it transition, and does it hold up to speed changes?

Or does it really matter? Who’s to say that a game animator NEEDS to build those blendtrees? Or that they have to understand how to set up an IK chain in-engine? Or know how the aiming system works?

In my opinion, animators will be far more successful if they have that knowledge.

I don't expect them to have it right off the bat, mind you. Entry level folks should be concentrating on creating good motion and letting their leads dictate the technical details. I also don't think it's a requirement to know scripting or understand how to build a complex state machine in order to be a successful game animator.


However, those animators at any level should be learning WHY those technical details exist. Why a pose needs to be oriented one way or the other, or how their motions will be used in game and why it works that way. They should know what scripting can do for them, at the least so they can request a tool from a tech artist. All the while, continuing to hone the creative side of their craft, in order to become the best game animator possible.

Marrying the creative side of animation with the technical side makes an animator even MORE creative. For example, on a project early in my career, I was tasked with giving the attack animations of a character “more flair.” This was all the direction I was given from my offsite corporate producer. I worked for weeks on just adjusting and presenting the animations with no success. By shear “he was a close friend” luck, I spoke with the VFX artist on the project on how we could push the attack animations further. He told me that I could attach VFX to my animations! I immediately added ridiculously exaggerated spinning weapon movements to the attacks, and worked with that VFX artist to create trails for the weapons. We hooked it up in game, showed it to our producer a week later and he exclaimed “that’s it!” Had I not known about that ability to tie VFX to an animation, I may never have succeeded in bringing my producer’s vision to life.

(at 3:16) More flair, Taurgis, MORE FLAIR! :)

Understanding what happens after your animation is done and makes its way into the game will allow an animator to devise new and interesting ways to make a character move. It also allows them to do it more efficiently, leaving more time for polishing animation or sneaking in that fidget animation that everyone keeps saying they should do.

Understanding both sides also creates the opportunity to be creative with the most tedious, technical animations, like those aforementioned transition animations. Mike Jungbluth does an excellent job discussing how to inject more character into those animations in a separate post. By not treating those animations as a technical requirement and instead as an opportunity to communicate a character’s personality, they become a creative necessity, not a technical one.

Use of motion capture is another opportunity to push beyond the technical. If treated like nothing more than technical data, the motion will stay as it was captured, and the performance delivered will fall short, no matter how good the actor (and we all know my opinion on THAT). Inject the creativity of an animator, and you have the opportunity to punch out poses, adjust timing for bigger impact, and even change shots after the fact. Yes, this opens up the "ownership" debate, but it’s the end result that matters. It’s important to remember that it’s a creative, collaborative effort, and not a technical hurdle that gets in the way of a performance.

Well, usually. :)

In the end, it’s best to approach both the creative and technical sides as wholly creative. The technical side varies from studio to studio, engine to engine, and it's ultimately a teachable process. The creative part comes from you, and you take that to whatever studio you work at and use it in whatever engine you work with.