Looking at the magic picture, am I right in seeing that raindrop splashes are affected by the topography of the ground, or are the animations just selected at random?
They are random right now, but having them affected by topology is a way cool idea. I’ll have to try that out and see how it looks!
Sometimes, fish swim in the water like they’re supposed to. Maybe they nibble at your toes, but they’re generally well behaved.
OTHER times they float eerily in the sky and slide toward you without moving their bodies. You can run, sure, but you can never forget that dead look in their eyes as their ever-gaping mouths follow you slowly, slowly.
Below, pie charts of of the hues found in various parts of my game.
I guess I like blue?
As an aspiring game developer, I couldn’t help but notice the procedural music in your game, such as the notes being played when the player hops on those flower buds. It’s a very powerful idea that never occurred to me, that the music could respond to the player’s actions in a very satisfying and polished way. Could you talk a bit how you accomplish this effect, what music theory you incorporate in its execution and how what parameters you vary to achieve certain “moods” for each song? Thanks!
For those who haven’t seen it yet, these flowers can be heard in action here. The video is of the IGF showfloor build, and features me stuttering and trying to answer questions about my game and saying ‘uh…’ a lot.
Let’s talk procedural music!
I love this stuff, and I believe that games, as a non-fixed media, are in a unique position to use it. Digital games are built out of procedure, it makes sense that their music should follow. There are a couple different ways to approach this when you’re making a game, but, for simplicity’s sake, I want to focus on three main modes of generation (which are by no means comprehensive). The first we’ll call procedural composition, the second procedural improvisation, and the third procedural mixing.
Procedural composition is concerned with creating music at a more granular level. One of the foremost examples of such a system is probably Emily Howell. It’s easy to think that if we’re making procedural music that this is the kind of thing we should be striving for, something totally computer generated. However, while this music is interesting for its own sake, it is always more or less composed in a way that draws attention to itself, and disregards the actual object it is scoring (although this is not a universal phenomenon). I’m trying to think of examples of games which take this approach, but it’s difficult. Disasterpeace’s January might be one example, although one could make the case it really belongs in the procedural improvisation category (and certainly, it can be said that the game exists in service of the music, rather than the music in service of the game).
Procedural improvisation takes a preexisting musical track and simply improvises on top of it. This is my favorite approach for a couple of reasons. It’s easy to implement and it can reinforce what’s happening onscreen especially well. This is the approach I used in the example you mention, with the flowers. A drone plays throughout the level, establishing a key for the flowers to play against. Whenever the player lands on a flower a random note from a harmonious scale is played. The computer also checks when the note is played that the same note was not one of the last two notes played. This creates a meandering kind of melody that isn’t necessarily concerned with the rhythm or overall composition of the music. More or less, I give the computer a set of musical boundaries to work within, and it creates something at random based around those. Procedural improvisation is an idea I’ve returned to over and over in my games, from some of my earliest works, to recent small experimental projects. Central to this mode of working, I think, is the idea of musical play and game play being combined into one object. One might think of it more as ensuring that all of your sound effects are explicitly musical. Although, to be fair, such improvisation doesn’t need to be tied to player actions.
Procedural mixing is where all tracks of music are precomposed, but an algorithm decides when and at what volume they play. Proteus is probably the best example of this used in a game - the volume of a track at any given time is determined by where and when you are in the gamespace.
Will the levels be separate entities, such as in Super Meat Boy, or will they be more continuous, like the finishing door brings you immediately to the entrance of the next level and have chapters of levels?
A bit of both, let me explain.
In videogames, we can understand the ‘level’ as a kind of vessel for an idea. Very focused, small levels often contain a singular idea about the game’s play. Because, for whatever reason, it’s the only game coming to mind right now, I’m going to use Give Up Robot as an example. However, it’s a pretty good example as most (though not all) of the levels fit into a single screen and contain a single idea. Ie: the second level’s idea is about how it feels to release from a swing and land on the ground - the third level’s idea is about how it feels to begin a swing while already falling - and so on.
The level in videogames, then, is not so much an object itself as it is a method for organizing information. Very focused levels contain single ideas, but this needn’t always be the case. We can imagine on one end of the spectrum a game’s idea parceled out each to one level, and on the other end we can imagine a game whose ideas are all present in a single, continuous level. Consider the difference between Give Up Robot and Limbo. (fun fact: Limbo, despite being continuous, still neatly divides its ideas through moments of stillness, or what we might call negative playspace)
In between these extremes fall games like Zelda or Fez.
With TFIJ, I’m creating ‘sets’ of levels. The levels inside sets all connect to each other (when you exit one, you enter another so that it is always possible to move through them in reverse order as well) and are each grouped around one particular idea. The ‘day’ set, for example, is grouped around ideas about making the player more comfortable on the jelly surface (the day levels are the first set, thus). The ‘night’ set is grouped around the idea of putting those lessons learned in the day levels to the test, and making sure the player really knows them before more complex ideas are introduced.
The level sets connect to each other horizontally (one set leads into another, and, like the levels themselves, one may move through them backwards or forwards). Each set serves as a way of dividing certain kinds of ideas in the same way that a level divides a single idea. There are also sets that exist outside these main sets. One set of levels sits ‘above’ the others, acting as a set of level sets. Other sets exist ‘below’ the main level sets (ie: secret levels).
I want the connections between these to be continuous, although the levels themselves are sharply divided.
So, uh… to answer your question…
Levels are divided much like Super Meat Boy, but I’m doing away with a menu for navigating between them. I didn’t like how replaceable the levels in Super Meat Boy felt, they had no relationship to each other (apart from visual themes or themes like ‘levels with missiles’). While my structuring is similar, I’m trying to think of the levels as continuous and related despite their separation. Each level should feel like part of a place.
Hm… I’m still struggling to find the words to describe it. Let’s jump to food metaphors.
Super Meat Boy levels are like chips in a bag. Although they bear resemblance to one another, the chips are basically disconnected. I want TFIJ to be more like a pie. You can cut out pieces from it and they stand on their own, but the organization of the contents in that slice of pie is influenced and directly affected by its mother object: the pie as a whole.
Wait - that made no sense.
It’s time for a look at… The Official Haphazardly Scribbled Notes of The Floor is Jelly (!)
Math. (Detailing formulas for calculating a quantized x and y distance in a single number rather than two. Ended up being not so useful.)
Early catflower sketches.
“midnight / longboarding with / a cigar is my idea / of a good time” I don’t remember writing this.
Are you going to be releasing it on Steam/Greenlight?
I’d love to put The Floor is Jelly on Steam.
Let’s talk about Greenlight.
Here’s a list of awesome games that, as of writing, Greenlight has excluded from access to Steam:
Thanks to Greenlight, as it currently operates, I would imagine that most of these games will never be on Steam. These are good games. These are games that deserve to be on Steam. If these can’t make it, what chance do I have?
In addition, and this is an old point to those familiar with the Greenlight controversy, the fee is too high for me. I live on a budget of a few thousand dollars per year. To me, $100 is a non-trivial amount of cash. Ordinarily, that money might go to living expenses, or to dealing with an unpredicted crisises (broken equipment, doctor’s visits, etc). Spending $100 on Greenlight, I may as well be throwing that money away. I’m better off trying to invest that fee in my game in a way that will actually make a difference. So, even if I were able to borrow that amount of money from friends or family (I’m fortunate enough to be in a situation where this is possible, but not everyone is) I would feel irresponsible using that money for something I believe will come to nothing.
My main concern with Greenlight is this: although it is made to seem a platform of inclusion (all the verbiage on the Steam page speaks of it in terms of putting games on Steam), it is rather quite the opposite. Only 40 games have been Greenlit so far (of which only 12 are available for purchase), but countless others never will. This is not because these are games won’t sell well on Steam, rather, this is because these games happened to be lower than the top ten on Greenlight at whatever point Valve decides it wants more games. Greenlight is a platform of exclusion. Although Valve advertises it as a way to put games on Steam, it seems to have restricted the influx of content to be less frequent and less diverse. Greenlight would be wonderful if Valve used it as a tool for finding games which have any audience to sell to, but it hasn’t worked that way so far, and honestly that’s baffling to me.
That said, perhaps this situation will change in the future; I am hopeful that it will. I’m not opposed to putting The Floor is Jelly on Greenlight, even if it may seem that way from this post. I only want to put it there once the game is done and ready for sale, however. I’m simply not comfortable putting the game on Greenlight while it’s in an unfinished state.
tl;dr: I hope so.
I’ve put a lot of words on this blog lately. Words are boring! I’ll give you all some nice pictures next time.
Page 1 of 6