I’ve seen a few times here in the forum about the performance of tiled sprites and that they need a lot of processing and therefore aren’t really performant.
I tried the VictrisGames Load testing example (fascinating) Load testing - a game example from the GDevelop game making app | GDevelop and the conclusions from that seem clear. That is - many normal sprites will perform much better than many tiled sprites. For example, 500 normal sprites run fine but 500 tiled sprites slow things down a lot to less than 10 fps. And of course it gets worse with a thousand or more objects. The sprites and tile sprites in the example are 6 by 6 pixels.
While I accept that tiled sprites require extra processing, I feel that it doesn’t really make sense to compare them one for one with normal sprites. If we are just using them as a static repeating pattern background then one tiled sprite can replace hundreds of normal sprite tiles placed/created in a grid, or replace one gigantic normal sprite.
I modified the load test example to show 10 objects with an increase increment of 10. It wasn’t until I had 40 tiled sprites that there was any performance drop and that was to about mid to high 50 fps. I realise that there are other factors to consider like object culling; the load test project isn’t set up for that.
But anyway, I’ve always avoided using tiled sprites but now, for me, it’s time to give them a go.