When opening the Emitter (DX11.Particles.Emitter Layer) help patch the particle count with the default 200 x 200 resolution is at 347200.
Shouldn’t it be 200 * 200 = 40,000 instead?
I was also wondering if only filled parts of the layer contribute to the particle number or the particles that are calculated. Does a larger empty/unfilled area result in less active particles thus better performance compared to a completely filled layer?
It appears, any emitter will add to the default particle system no matter if connected to this particle system with a wire or not. They can be separated by adding a different particle system node and change its “ParticleSystem Name” pin as well as the “ParticleSystem” pin of the emitter to a different name.
The particle amounts are correct now.
@io sounds kind of similar to what you were saying, now that I’m reading your post again.
Well, that’s the way it works, the particle system name is the name space where particles and behaviours are attached to (so to say, no need to create a vvvv connection, the name space is the connection in memory). As things are now, when changing names and reorganizing your systems, just remember to close everything and reopen.
I would like to get back to the second question of my initial post:
I now know all particles will be in the buffer no matter if visible or not. But i was wondering if it was possible to filter the buffer for the visible particles only? What selector would I have to choose for example?
Particles work as an append buffer, thay are filtered under the hood, however buffer count = element count even tho slices are discarded.
You can get real count with a copycounter node, hower to do that you must read back from gpu wich is slow expensive call
There is a buffer called AliveIndexBuffer that holds all ids to particles that are alive in the current frame. In addition the first value in AliveCounterBuffer gives you the number of alive particles. You need this value to adress the AliveIndexBuffer: AliveIndexBuffer to AliveIndexBuffer[n] where n = AliveCounterBuffer gives you all pointers to living particles in the particlebuffer.
There is no need to do expensive readbacks. You can access these buffers everywhere with computeshaders. Have a look at existing nodes. There is a difference between the adressing before and after rendering the buffers in the particlesystem node. have a look at modifiers (=before) or AsAttributeBuffer (=after).