Add amount_ratio to cpu particles to dynamically change amount emitted

[Ducklett]

Revival of #79527

This PR implements the amount_ratio field for CPUParticles2D and CPUParticles3D. This already landed for GPUParticles2D and GPUParticles3D back in Godot v4.2 but the original PR for CPU particles never made it in.

Changes from the original PR

• changing amount_ratio will no longer change the emitting flag
• updated docs for amount and amount_ratio to match those for GPUParticles2D and GPUParticles3D

Fixes godotengine/godot-proposals#5939

[arkology]

This PR suppresses #70145

[Mickeon]

This PR suppresses #70145

Supersedes* 😅

[KoBeWi]

I think the OP has meant to mention it instead of #79527

[Ducklett]

I think the OP has meant to mention it instead of #79527

Yeah, my bad 😅

[KoBeWi]

That last change seems to simplify the code a lot.
But it doesn't work 🙃
When you decrease ratio, the particles freeze instead of disappearing.

[Ducklett]

When you decrease ratio, the particles freeze instead of disappearing.

I see, it's keeping inactive particles in their final state and the effects I was testing with had a scale/fade out so they seemed to be gone. I'll look into it.

[Ducklett]

When you decrease ratio, the particles freeze instead of disappearing.

Solved this in the latest commit.

[KoBeWi]

The same condition is few lines below.

[QbieShay]

This needs testing to check how it behaves compared to GPUParticles. Code here:

godot/scene/resources/particle_process_material.cpp

Line 854 in c5c1cd4

code += " if (rand_from_seed(alt_seed) > AMOUNT_RATIO) {\n";

GPUParticles set particles inactive at birth depending on a random value

godot/scene/resources/particle_process_material.cpp

Line 854 in c5c1cd4

code += " if (rand_from_seed(alt_seed) > AMOUNT_RATIO) {\n";

This gives it an organic feel while still maintaining that zero has no emission and 1 is full emission. Your PR kills one particle every X, which is not the same behaviour.

[Calinou]

This gives it an organic feel while still maintaining that zero has no emission and 1 is full emission. Your PR kills one particle every X, which is not the same behaviour.

See discussion in #104276. The "kill one particle every X" behavior seems more desirable overall, especially at low particle counts.

That said, we should aim for consistency between CPUParticles and GPUParticles, so this PR should be changed to use the random approach for now.