Crab Meat: Pivoting for improvements
I’ve been working on Crab Meat for not much longer than a fortnight now and there are concepts, ideas and implementation that I’ve had to change or pivot away from to better the game.
Camera Movement
Originally, the camera was focused on the player. The camera was zoomed in and whenever the player moved, so did the camera. However, I realised that players weren’t able to plan a strategy and that we’d either have to make levels larger and zoom the camera out at cost to performance, visibility of each block as well as spend more time designing the levels.
As zooming out meant more for the camera to render and would take longer to design a good puzzle, I convinced the other members of the Stern Buoys to use a fixed camera with a view of the entire map. However, this means that a levels can only be a certain size so that the player can still see the screen but now the player is able to plan a strategy or path for themself.
Level Sizing and Design
Originally, the size of the level could be what ever I wanted due to how the camera worked, but because the camera became fixed, I had to also limit the size of the levels. Before the levels went from essentially being infinite across and down, changed to 20 down and 15 across. This became the perfect size for a level, as it was easy to see every asset as a player and every entity became visible on screen.
I first had it that every level had multiple screens, so a player could enter a tunnel and it’d become a new screen, even though it was part of the same level. This became too much for the team to implement in the short period of time for Crab Meats prototype so it was changed to be that each room is it’s own level.
Post Effects
As mentioned in another developer diary, I had multiple post effects (Ambient Occlusion, Sun Shafts, Motion Blur and Depth of Field) to try and make the project look as appealing as possible. However, the fact that my personal computer is 100% more likely to be far more capable of running post effects than an Android Tablet had escaped me while implementing them and I realised that I had to take most of them out.

I first assed which post effects made the biggest visual change and which ones caused the most stress. I came to the conclusion that it’d be best to remove everything besides Ambient Occlusion for optimisation purposes.