Archive for the ‘Behind the scenes’ Category

Designing for Action vs. Puzzle


2011
04.03

A ‘behind the scenes’ look at Confetti Carnival’s development process

In this post I’ll take a closer look at the actual gameplay mechanics (with some new footage) of Confetti Carnival and share my thoughts on an issue that we’re dealing with for a long time.

I’ll start by teasing with a little video showing 5 different ways of getting bombs in Confetti Carnival. We’ll get back to that video later on so you might wanna watch it again after your done reading.

Watch HD version on Youtube

A little background on the game

Before I go into things, I need a few words to explain the core concepts of the game to those who haven’t played it yet (which is almost everyone, since the game is still in development).

In Confetti Carnival, each round (of each level) introduces a group of blobs called Splatters and a set of Confetti Bombs. Your goal is destroy all the Confetti bombs on stage using the Splatters your given to play with. The way you destroy the bombs involve flinging your Splatters onto walls and watching them break up and spill their liquid. When bombs are washed by liquid they ignite and explode into confetti.

You play each Splatter at a time. You just point in the right direction and hit the ‘Slam’ button – the rest of the action is performed by physics simulation and learning to anticipate it is part of the skill the player needs to hone.
As the simulation unfolds the player can intervene at any time by hitting the ‘Flip’ button. What it does is reverse the motion of the entire scene (Splatters, bombs, liquid … everything!). This, turns out, is a very powerful and yet simple way of manipulating the simulation and players will consistently find new scenarios where it comes in handy. On top of that there is a cool (in my opinion) stunt and combo system that rewards shots that are more challenging, stylish or otherwise impressive.

Designing for different play-styles

Our design was aimed to enable players to ‘choose’ how intensely they want to play it. They can take their time carefully considering each shot or just go ballistic on it and try to clear the entire round (or even an entire level) with a very long combo of moves. This was mostly achieved by two main design decisions:

  1. There is no time limit on rounds/levels at all.
  2. If you wait for too long between stunts (~3 seconds) the current stunt combo will end (and score multiplier resets).

Pretty standard stuff, but what we are seeing so far, is that most players start by playing it slowly, which is expected since the game has quite original game mechanics. What we’re trying to do now, is slowly push players towards playing faster and doing longer stunt combos. When you try and achieve long combo strings, you need to play faster and adopt a more ‘reactive’ play-style where you roughly aim your shot and then use the ‘Flip’ and ’2nd shot’ mechanics to correct your move as needed. But that’s not the whole story – the rational behind our combo system is to deliberately make shots to generate more stunts even if the ‘easiest’ way to clear the round is more direct.

… Continue reading …

Rendering without meshes


2010
06.16

Well, I’m actually showing off some work that was done a while back – but it’s still nice.

What you see here is an early version of one of our game creatures (aka Gummy) jumping around on a small part of a test level. The interesting thing here, is that no meshes (you know vertices, polygons and all that stuff) are used for rendering – it’s all done with sprites and some shading tricks.

Why I like particles?

To explain how it’s done I need to provide some additional background first. Our game-play is 2D in nature and is fully physics based. I have done plenty of physically based simulations for non-game applications, so when we decided to start developing Confetti Carnival it was a natural to ‘tap’ into that experience and to create our creatures from a set of particles (all particles have masses associated, embedded force elements and so on).  I won’t go into the simulation part right now, although there are interesting stuff there too.

The common practice in these cases is to create a mesh that can be ‘skinned’ over the elastic creatures. However, in this case, our creatures are made of a liquid/gel like substance, so I decided to explore ways to render them using liquid-like techniques.

Creating a surface using particles

Instead of generating a mesh, I render sprite particles with color and depth information onto a render target. The particles are created to cover the entire creature and depth is accumulated into the render-target. Each sprite uses a simple soft sphere-like 2D texture and the accumulated spheres gives the final shape. In the final pass, the render target is sampled to provide depth, color and normal at each pixel, which gives enough information to visualize it.

There are several great benefits from using this technique, especially in the context of what we are trying to achieve. But the most important and immediate benefit is that even when the creature is severely deformed, the resulting shape remains totally smooth.

What you don’t see in the video (but is already implemented and working), is the ability to use textures to give the Gummy a not-so-smooth look and also to add local creature details such as eyes, face and any other features the Gummy might have…

Hope you enjoyed this,

Niv