How I tackled my performance issues developing an Android game in Unity | Hacker News Reader

How I tackled my performance issues developing an Android game in Unity

codeandux.com

12 pointszerratar11 comments

I will never understand why people use Unity for simple 2d games when there are engines for 2d games that are just as good. They may not have all the bells and whistles, but you don't need anything but sprites for 2d.

Hey dubcanada :-) Hehehe. Unity is still a very nice game engine for developing 2d games. However, even if the game I created seemed to be in 2d. It was actually in 3d but with a orthographic camera. Since my intentions later on were to enable the players to switch between a 3d and 2d mode ingame.

Thank you for your reply!

It's quite simple. Nowadays you don't gain any performance by eschewing 3d engines, since even hardware-accelerated 2d engines are just wrappers around 3d hardware accelerators.

What specific engine would you have used?

Because a game engine isn't just a rendering engine, and Unity has a ton of useful features that enable rapid full featured game creation. "...you don't need anything but sprites for 2d" doesn't take in to account (and this is by no means an exhaustive list) all the things Unity and other game engines like Unreal offer that makes game creation easier:

- A game object management system

- A component system

- AI behaviors

- Sequencing

- Animation

- Pathfinding

- Physics

- User interface

- Resource management

- Networking

- Input mapping

- Asset importing and manipulation

- Asset optimization

- Audio

- An editing environment that allows rapid development in terms of asset management and live modification, as well as easy customization of the asset pipeline and game specific editing controls

- Easy deployment

Not to mention, a "2D" engine nowadays is much more than a sprite management system, they are written for 3D API's to offer fast blending, shaders, compositing, effects, transforms, etc, etc.

These aren't bells and whistles, they are tools that have been developed over years to work well for game production, and they are features that even simple games can take advantage of to make them more quickly and more easily add features over time.

Not to mention having a community of other game developers who use the same platform and a tool like the asset store which frees you from having to be an artist to make your game look good.

Unity doesn't provide any sequencing for you (maybe you were thinking of Unreal's Matinee?). Also, most people who need input mapping don't use the builtin input mapper. Editor scripts are also painful to write.

You're not really wrong. The platform support and asset importing and hot reloading, for example, are great. However, some of the things like resource management and PhysX or Box2D integration aren't that big of a deal. I want to be realistic about the actual value that Unity adds.

time wise it adds the collision engine, the ease of compilation into Android Ios with a few clicks, and plus all the engine for animating sprites and viewing the changes live, the sound engine, the UI tools.

The goal of using unity for 2d projects is because it is faster, more user friendly and easier to work in team with designers than using pure open gl and libraries.

Yes, the platform support and hot reloading are real nice. That's why so many teams have invested in Unity.

The collision engine is what I was referring to when I said it's not that huge of a thing to do your own Box2D integration, so I disagree that there's a huge value add there. Just drop in the same solution they dropped in.

And the sound tools, until very recently, were incredibly rudimentary. You'd have been just as well off tossing in OpenALSoft or rolling your own FMOD integration (at least then you could use the FMOD Studio tool). I'm grateful for the new Audio Mixer window.

Also, 30mb.

Please suggest 2d engine as nice as Unity ;-)

Don't see the point of doing such a simple game in Unity.. Phaser would probably be better..

You're not wrong, using simpler 2d tech would have made the process a lot easier, but this way it was a good intro and learning experience for making something bigger in 3d which will eventually require a lot of the same optimisations (and a lot more).

The only basic part missing in his article would be along the lines of combining textures into an atlas and dynamically modifying UV's so that for example the UFO's could have had different shapes and colours.

Reply on news.ycombinator.com