Justine Hamer- GDC Writeup

Post date: Mar 13, 2015 4:23:49 AM

Alright, so the Game Developers Conference (GDC) of 2015 was last week and I attended for most of the week. This is my second time attending the conference, so I felt a bit more confident about what I wanted to get from the conference and what I wanted to learn more about. This time around I also have an internship lined up, so I spent less time in the Career Centre and more time on the Expo Floor and in talks. So let's get right into it!

The first talk I went to wasn't technically a 'talk,' but rather a presentation at the Ubisoft booth about their new 'indie-style' game, Grow Home. Grow Home started as a prototype from a game jam that eventually got absorbed into Ubisoft proper and became a pet project of the company. It was a small team, so of course everything was procedurally generated, from the landscape to the animation, to minimise the pressure on the one artist on the team. It was interesting to see how they approached the gameplay: the way the world grew around the player, and the way the player grew the world around themselves, practically forces the player to explore this simple yet fleshed-out world. It seemed like an extreme version of optimization: no frills, just problem solving. Still, it feels like an Ubisoft game: you're given an overall arching goal, but only really as something to aim towards right at the end of the experience. Everything else is dedicated to exploration and side quests. It was also interesting to hear how it developed: as it was viewed mostly as an experiment, not a real game to be marketed, the developers could feel like they could take risks and iterate as much as they desired.

First proper talk I went to was Proceduralism and Automation by Luiz Kruel of Certain Affinity. Hey, guess what! You know all those Maya Python scripting tutorials? He did those! His voice is so soothing. Anyway, his whole talk was about how, as a tech artist, it's his job to ensure that, instead of hiring more people, he can make the people already there more efficient. He's in charge of developing a co-dev pipeline: a pipeline to be used by a co-dev studio, essentially a satellite studio embedded in the tools of the original company. Certain Affinitiy, as a co-dev company, has cool new tools, quick projects, and no downtime. These, Kruel affirms, are both pros and cons. We have to protect the artists from burnouts, multiple workflows, and wasted efforts. But how do we do that? With automated, symbiotic tools: for instance, the one-button exporter tool! This allows the artist to export their concept from zBrush, poly reduce it, bake it down, compose maps, assign material sets, and import it into engine in a matter of minutes. Artists can now iterate on their concepts in a nondisruptive manner, and with a databank of master materials made beforehand, all the materials on a model can look like they were made by the best texturer! Brilliant! Kruel also talked about how his team created a bunch of building blocks in multiple sizes to procedurally generate new level designs: plug these pieces into Houdini, and suddenly a block can be 'stretchy,' with caps and a middle. With this procedural set of prefabs and a simple blockout, suddenly worldbuilding has been automated.

A final parting note from Kruel was that, of course, some artists may balk at using these new tools. The best way to get them used to it is to have them use it, he affirms. That way they can see how it's meant to improve their workflow.

Second talk was How I Made My Game No Fun, by game evangelist Mike Hines. He wanted to make a popular, successful game, but he wasn't sure how to go about it. He started off by collecting a list of top ranked, top downloaded games from Amazon, talked to game developers, critics, and players to figure out what made a good game. Of course, he ended up with a really big list, so he figured he'd find out for himself. Thus was born Rock, Paper, Scissors Tag. Studies and tests suggested this original game was fun, but not that fun especially if you weren't very good at it. Thus comes the idea of 'rubberbanding,' 'game gravity,' or as I like to call it, the 'Mario Kart Theory of Economics.' This is a gameplay aspect which assists the loser when they're falling behind. This could make it more fun for the loser, but could also frustrate the winner. When it's added to Rock, Paper, Scissors Tag, both the winner and the loser have no fun. So how do we have fun? Hines posits we have fun when we get fully engrossed in a game; we shouldn't implement features, but feelings. We need to find some way to implement Flow so the player has longer sessions, more sessions, and a deeper engagement with the game. Jenova Chen made a full study of this Flow in his game, Flow. To create Flow, we need a distinct set of goals that have impact, few distractions with no arbitrary rules or irrelevant feedback, immediate and timely feedback, and a balanced sense of perceived difficulty. This is how you can make a game fun.

Third talk, perhaps a bit too dense for me, was Defense Against the Dark Arts: Budget and Budget Recording by Paul Vosper of Bungie. Another technical artist, Vosper advocated strictly for budgeting in games because when they shipped the first Halo game they had to cut down the texture sizes in every environment by three quarters so it would fit on the console. To mitigate this, following games like Destiny had budgeting shoed in by preproduction. Technical artists must wear many hats, Vosper adds. They need to be part oracle, part accountant, part detective, part human UI, and part educator. They need to create a balance of values between memory, gameplay, and visual aspects: the game must play, look good, and be fun. World art is easy to keep track of and budget, as it tends to be the first in and doesn't change much. Following that is sandbox, like character and vehicles, and finally design comes in last. There are resources linked to the cost of removing content: cutting assets is the most pragmatic, and therefore most important, action. Tools must be made to track budgeting; to get data, but also to understand it. These need to record what happens at runtime, as this is the most important. Programmes like Maya and Max read vertices differently to engines, as engines also count skin weights and other things. So a vertex estimator tool is also handy, as it dispels this illusion. Most importantly, budgeting needs to connect cause and effect: curators and creators need to work together to track down issues and resolve them. In the end, a web page is made to track trouble and how they change between builds. This way, the Dark Arts can be vanquished!

Fourth talk was an exercise in biting off more than I can chew. I attended a talk by Intel and Blizzard about Edge Detection Based Post Processing in Warlords of Draenor, presented by Matthew M. Williams and John Hartwig. I kind of thought I would be able to understand post processing, as I had toyed around with a lot of the processes in UE4, but I guess I had underestimated just how much of an artist's tool Unreal Engine is. As soon as they got to the slide-long equations I realised post processing was a far more complicated process than I had originally thought. I left the talk with a better understanding of how lucky I was to work with the Unreal Engine, as well as a better insight into the thought processes behind hard wiring post processes in in-house engines. Seeing them take apart the sketch shader was really interesting, and I'd like to see if I can at least apply some of that knowledge into my future attempts at post process volumes.

On the same day as the fourth talk, I got to visit the WB San Francisco headquarters and get a tour of the facilities! The offices are in the same building as Flickster, as this arm of WB is relatively new, but once the studio has gotten a few games out it should move to somewhere a bit more suitable to the WB name. The offices themselves are nice and spacious, with a lot of memorabilia from different WB IPs all around. I'm not sure how much of the visit I'm allowed to talk about, as I saw some of the prototypes for the games I'll be working on during the summer and it's kind of a secret. I can only say they're going to be mobile games, as AAA studios are considering muscling in on that industry. It's a pretty smart move, as more people have smart phones than any console whatsoever so it's a far larger market to have access to. I'm very excited to start working there in June! I also got a super fancy hoodie:

The next talk I went to is one of my personal favourites; I went to a similar one last year and loved it. Yes, it's the 10 Microtalks, 10 Speakers session! Each speaker has 5 minutes and 22 seconds to talk about whatever they feel is important to them, no holds barred. First up was Richard Lemarchand, who introduced and MC'd the whole discussion.

Then came Emily Short, who wants protagonists to have more agency and character than they do. Antagonists are always so fleshed out and involved-- why must protagonists always be such a tabula rasa? She cites particular RPG games like Fiasco, Polaris, and Monster Hearts as games were the development of the protagonists' character at the hands of the player serves as the most important part of the gameplay. History can serve as a currency, she says-- for instance, if you seduce someone, that should affect how susceptible they are to your suggestions later on. Have characters that remember you.

Lisa Brown followed with a talk about efficiency versus accountability. When working on her game, she tried a number of things to keep efficiency up and accountability related to that. She kept a design log, essentially a blog, and found that shaped the design process due to performance pressure. She felt there needed to be something to account for every log and a way to justify each practice. A second thing were the developer streams, which were less academic but helped build a community. This made the team sensitive to visual progress, making sure they looked like they were getting things done to outside viewers. Finally, she also tried weekly builds with hyperspecific checklists-- with specific deadlines, you can close design loops and commit to decisions. All of these things, she decided, made you accountable to your decisions, but more importantly they made you accountable to people you care about, like your coworkers and game community.

Then came Matt Boch, who wanted to talk about empathy with characters, first mentioning Tinder as a great free-to-play mobile app. He advocated for more ways to battle the hyperentitlement of players when it came to romancing NPCs, specifically citing Dragon Age: Inquisition and how many people rerolled as a female so they could romance the NPC Sera. This creates a relationship where the player is blind to the desires of the 'goal,' the person they wish to attain. To fight this, NPCs should have more random actions, more 'romance permadeath,' so the player has to think about the character as, well, a person as opposed to an object.

Naomi Clark championed broken games-- what, in actuality, are broken games? If games are made to 'solve problems,' then what problems should be solved? If it's boredom, than games just become a means to an end as opposed to an artistic statement. We need to view our games as creatures being gestated: sometimes glitches will develop, but a little bit of broken will go a long way in creating an extra level of unpredictability and joy in development and play.

Tim Rogers gave a talk, but I couldn't find much merit in it. He just talked about himself in relation to games journalism and patted himself on the back for calling Bioshock bad, or something. I don't know.

Holly Gramazio game jammed with us, giving us such delightful game possibilities as Cars = Monsters (don't let the cars 'see' you), Step in Shadows Tag, 50/50 Moral Dilemma (try to divide your friends up by morals as neatly as possible), Joust with Ice Cream, Tiniest Dog (I can't remember what this was, but it sounds awesome), Rotate your Sleeping Partner, Christmas Tree in New York, and Stone. There were a lot more, but it really just went to show how easy it is to play without really thinking too hard about it.

Celia Pearce schooled us thoroughly on the Fluxus and Dada movement and how they contributed to the development of play as we know it. Duchamp made games, especially chess sets. Elsa von Freitag made readymades to be interacted with. Dada work was meant to be played with, not just seen, so the design can be fully experienced. Rules, randomness, and chance were huge aspects. Fluxus made toys and the first game mods like Fluxus ping pong. It always helps to reference history when making games.

Cara Ellison posited that games don't give us as much intimacy as they should. We need intersubjective intimacy, like when in Out of This World the alien touches your face. The NPCs in the game need not to be others, but real. There needs to be shared needs, and wants-- emotional attachments through accidents. The player needs to be treated as someone emotionally intelligent.

Finally, Rami Ismael tore the games industry a new one for not having enough of an infrastructure to support any other language but English. We make games where we go into others countries to shoot them, but we can't get their language right. Non-English speakers have become the invisible majority-- do we have a programming language for Arabic? Why are all tools and documentation in English? We need to make more efforts for diversity of all sorts.

Compared to last year, this year's 10 Microtalks series seems to focus more on empathy and exploratory play-- interesting, considering it's the first Microtalks series following the Gamergate issue.

I also went to The Art of Monument Valley by Ken Wong of ustwo games, which is a smaller part of a larger company that doesn't actually produce games. Monument Valley grew from art and story, not gameplay: as opposed to having a prototype and asking 'how to we fill this with art?' the team had a specific vision in mind and asked 'how do we make this a game?' By placing the visuals first, the game naturally blossomed into something unique, and by having unique visuals, marketing was far easier.

The Ulllltimate TRIM!

Finally I went to a talk about The Ultimate Trim in Sunset Overdrive, which was about how they texture packed with a standardized UV layout and a specific normal map with 45 degree bevels, but I think I've used this technique before with some of my handpainted stuff when I didn't want to overuse UV space. Still, the results looked very cool and allowed for a very smooth game. By having standardized UV layouts, they were also able to make Maya tools around this for unwrapping. Overall, this process saves time, cuts down on lowpoly edges, is programme agnostic, and helped automate the whole process. Efficiency!

Overall this GDC seemed to focus more on efficiency and empathy, things we all need to work on. I was surprised there weren't more advocacy talks, considering how quickly this industry is changing in regards to diversity and inclusivity. I know some people found there to be too much, but I think it's something that can't be talked about enough. There's no harm in letting more voices be heard, I reckon, from all walks of life so long as they're relevant to games.