2D Platformer: Tiles vs Complete Shapes

I’m a developer and this is the first game I will be making using unity. I would like to know if it’s better to ask my designer to create the assets as tiles or as complete shapes, giving that i have the entire design of the level set in mind, and could someone also point me to a tutorial on how to set-up unity so it’s tile map ready ? for example 2d maps, snapping…etc.

You need to understand WHY people use tiles. It’s not because it looks cool, usually. It is a form of compression. In order to store unique textures for every piece of the environment can consume a huge amount of memory, ie you don’t re-use any of it and you don’t compress any of the redundancy. This could never work well in the older computer systems because of major memory limits so you had to get clever. So if you could conceive of a similar enviornment composed of re-used pieces, you can make much better use of memory. And of course rectangles are the easiest shape to deal with so tilemaps started out based on rectangular tiles. Essentially you’re performing a compression algorithm where you’re saying this tile is the same as the tiles at all those other locations, so just store the index of the tile instead of the whole tile. Then displaying the tilemap is a form of decompression, taking the index of the tile and turning it into a preset block of pixels.

Obviously there are side-effects to this - your environment may end up looking blocky or rectangular-based, depending on the size of the tiles and the amount of variety. But with large enough and varied enough and clever combinations of tiles you can alleviate most of the impression that it’s tile based, especially if you don’t have tile blocks that obviously look like a block with a perimeter around it. A game like Rayman Legends for example does not look obviously like-based but it is. But then again, you might be going for that as an actual art style in order to look retro or something. There are other forms of compression too but tilemaps are quite easy to deal with. You need to decide how much re-use you want of resources, and remember it takes extra time to draw graphics, so if you are not re-using any of the graphics more than once that’s going to mean a heck of a lot of time is needed to generate it all, versus re-using blocks in different combinations in order to get a higher return on your investment.

Yes, the main reason I didn’t want to use tiles is because I don’t want my game to look ugly, I took a look at “Rayman: Fiesta Run” and I think they are using a mix of rectangle tiles, and other shapes, I also noticed that not all the shapes have the same size. What I’m leaning towards is to use different shapes with different sizes so the game doesn’t look dull or old school, and of course I will be re-using most of the assets, I’m just not sure if I have to go the traditional way of using those rectangle shapes. And wouldn’t the compressing - decompressing process be irrelevant if I use spritesheets ?

More modern 2d games bypass the need to work with a grid and instead allow you to place objects anywhere, which is really Unity’s native approach - just position your sprites wherever you want them in whatever rotation or scale you want them to build your level. There is no compression if you do not re-use a sprite, there is if you do.

Exactly. The main reason to do so now is keep the drawcall down.