Here I am actually talking about performance question.

I've already done some research about creating layers in html canvas projects, and I found in most tuts, they just told me to write a lot of static html canvas tags, then get context from each of them (then use css / absolution position/ z-index to simulate the order of layers).

But I am wondering if THIS WAY is better:

1. write only ONE canvas tag(and let's call it background canvas) in html.
2. create other canvas element using createElement api for each layer, but not appending them into DOM.
3. get context from these "virtual" canvas(which is dynamically created)
4. use canvas api to draw things on these "virtual" canvas.
5. use requestAnimationFrame and drawImage to draw every "virtual" canvas onto background canvas

Actually, I thought dynamically created canvas is more acceptable/flexible to me, but I don't have a lot of experience in canvas performance optimization (And it's why I am here).

I think I need some guide about how to test performance of this two cases.

If anyone here have any suggestion will be appreciated.