Games start with a GDD, or Game Design Document.  Games that do not start with a GDD are pretty much doomed to fail, because they will have no concrete guidance on what game they are making.  The GDD contains story and character information, but it also contains much more than that.  A GDD will describe what genre the game falls into.  It will describe gameplay decisions and the consequences of those decisions.  It will describe character movement, enemy types, weapons, and puzzles.  It will contain descriptions of the levels, and the actions the player will need to perform in order to progress through said levels.  A GDD is the game, albeit in a 2-d, non-interactive form.

GDDs are important for both the developer and the publisher.  They’re used to pitch games to publishers, so both parties understand exactly what kind of game is being discussed.  The GDD will of course be amended and altered over the course of the project, but it’s still a very vital starting point.

The GDD for Grim Fandango can actually be downloaded in PDF form from here.  It’s worth a read for aspiring designers.

Now that that is out of the way…

On a previous post, Robert asked:

So, I have a question pertaining to 3-D modeling.  For the purposes of my portfolio, I’m wondering how high-res my pieces should be?  I frequently find myself creating models that are shaped very well, but they have a rather large poly count. So, I guess the underlying question here is whether it is more impressive in a portfolio to see very high poly count models with good texture or to see lower poly count models with textures that just give the appearance of higher resolution?  (Keeping in mind that my painting/drawing skills are rather poor at present so the textures are a definite weakness for me right now)

This is a good question, and the answer is almost always, “It depends.”  As a general guideline, you should never put more geometry into a model than is absolutely necessary to create the shape you want.  If you’re making a cube, that cube only needs to be 6 polygons.  A cube that is 600 polygons is ridiculous, because you really only need 6 in order to make the cube look like a cube.  The more complex the object you’re creating, the more polygons you’ll probably need to make it look right.  If you have a set budget for polygons, use them where you’ll get the most bang for your buck.  People won’t care if a door is made up of six polygons, but they’ll notice if the doorknob is blocky and unrealistic.  Have a look at the wireframe on one of my meshes:

This is a fairly high-poly mesh, but I spent those polys where it would show.  If those curves were choppy and blocky, the mesh would just look bad.  Since I spent the polygons on those important parts, the blocky headboard and footboard can be very simple.

Generally, it’s more impressive to see a lower-poly model that looks good than a ridiculously high-poly model.  Anyone can throw extra polygons at a mesh to make it look smoother, but it’s much harder to serve modeling steak on a hamburger budget and it shows more skill.  Texturing is also really important–if you model really well but you can’t texture for poo, you’ll probably have a very hard time getting hired.  There really aren’t separate modeling and texturing positions in most companies–3-D artists do both, so good skills in both areas are pretty necessary for the job.

Thanks for the question!  I hope this was helpful.

Into this:

It’s about 22 yards of my very first handspun yarn, and I’m quite proud of my lumpy messed-up wool baby.  I’m finding spinning to be very relaxing, in a similar way to yoga.  You do some of the same movements over and over again, slowly and carefully, and just pay attention to what happens.  In yoga, what happens is you get some nice exercise and are hopefully a little closer to being able to touch your toes when you bend over.  In spinning, you end up with yarn!  I found it so enjoyable I felt the need to repeat the process with another pile of roving:

Mmm…   The colors.

PS:  If anyone has any questions they’d like to me answer about video games or whatever, please ask ‘em in the comments!  I’d like to make the Q&A a more regular part of the blog.

But mostly I’m excited about the new fridge.

Anyway, a few days ago commenter Chandler asked a couple of questions:

how much time would you say is spent testing the games you’re working on compared to actually working on modeling and stuff for them? What kinds of things do you look for specifically as a graphic designer when testing?

I’d also maybe like to hear more about the outsourcing of modeling you mentioned… how severe is it, and how do you feel about it?

How much time I end up testing really depends on how far along in the project we are.  If we’re still in full production, I don’t test really at all because I’m busy creating assets.  In the early stages, there’s not much point in playtesting because there are no completed levels.  The later it gets into production, the more levels there are to test and the more we’ll end up playtesting them.  Since the art tasks tend to be finished before the design tasks, it naturally falls to the art team to pick up more playtesting as the project progresses.

When I playtest, I’m actually not looking for things from an artist’s perspective!  I’m checking to make sure the game works.  We playtest to ensure the levels can be completed and that there are no major bugs that are either reeeaaally noticeable or stop the player from proceeding.  I’ll see a lot of art issues, of course, but that’s not the point of the playtest.  The point of playtesting is to make sure there are no places where the player gets stuck on collision, trapped in a wall or falls out of the world.

Outsourcing can be an issue, but not always.  The thing about outsourcing is that it can be really helpful in dynamically scaling your art team size.  You don’t usually save money on a per-asset basis, but you can essentially hire 30 extra artists for a month to crank out a bunch of stuff for your project, and then not have to find the money to keep those 30 extra artists around for several years.  Outsourcing is really only useful for characters and non-essential props for the most part.  If you try and outsource things like a wall system or a catwalk system you’ll end up wasting loads and loads of time getting the outsourcing team to tweak those assets to your specifications.  It’s much more efficient to keep important assets like that inside the company, where the turnaround time is 15 minutes.  The further outside the company you go, the longer the turnaround time is.  A 12-hour turnaround time for revisions really adds up if an asset needs more than one tweak!

The other thing about outsourcing is it needn’t be outside the country.  There are outsourcing firms in the US.  We sometimes work with a Seattle outsourcing company.  Outsourcing companies can be a great way for artists to get experience, too!  You might not get a job with a game studio, but working as a 3-d artist at an outsourcing company will get you into the industry.  I personally don’t feel threatened by outsourcing, because a studio is always going to need some in-house artists to ensure quality and quick turn-around.

Thanks for the questions, Chandler!