Drawing Buffer

[yahiko]

I would like to know if by default drawing graphics (PIXI.Graphics) with pixi.js implicitly uses a buffer or if we have to do that explicitly. If so, is there a way to do that natively with pixi.js?

For the moment, I have implemented a small class for this purpose, but maybe this feature already exists in the core library.

Thanks for your replies

[ivan.popelyshev]

What kind of buffer are you talking about?

[Taz]

Maybe cacheAsBitmap could serve your purpose if the issue is optimizing for things that don't change often?

[yahiko]

Let's say I want to animate 10k rectangles on the canvas.

I maybe wrong, but I would draw these rectangles on a buffer first before displaying it all at once, by switching the visible attribute of PIXI.Graphics. This is what I call a drawing buffer

[Exca]

Render your graphics into renderTexture and then use that as texture for 10k sprites.

[yahiko]

In fact, I don't really want to turn my 10k rectangles into sprites since they could change at every frame (shapes and colors).

[ivan.popelyshev]

Just take "Texture.WHITE" as texture, it will be 10x10 size, change width/height of every sprite (however i recommend to change scale instead). For color - change "sprite.tint".

[yahiko]

That's absolutely true, I know this feature in Phaser which is based on Pixi.js until v3, but I forgot to mention that my graphics' lifetimes (not only rectangle in fact) are limited, and I produce new ones constantly, like a particule system in fact.

Say differently, I really need to draw them programmatically at every frame.

[ivan.popelyshev]

Reuse old ones  Just dont screw up remove operation. Whether its just visible=false or remove from array, it can have O(N) complexity instead of O(1) per one removed element. That's O(N^2) if you remove everything and its really bad.

[yahiko]

Good and quick discussion. I am going to stay with my home made implementation, but I will keep in mind what you have said guys. That could be a way to optimize my project

[ivan.popelyshev]

18 hours ago, yahiko said:

Good and quick discussion. I am going to stay with my home made implementation, but I will keep in mind what you have said guys. That could be a way to optimize my project

Some people have inventors mind and find it difficult to use things they dont understand. Good luck to you!

[yahiko]

As a matter of fact, I keep my home made implementation for the moment because it is satisfaying enough, performance wise, tt is also because I need to finish more important stuffs before optimization, and because of some lazyness of course!

But what you and the others have written will surely inspire myself when I will be forced to optimize thouroughly with increasing objects and computations.