Example Maya scene of gasoline explosion

[dgould]

The reasoning for keeping the Max Step low and the SPF also low is to produce a slow motion simulation. When I changed the Time Scale the characteristics of the sim were very different. After testing it seems that only the Max Step and SPF are the ways of slowing down the sim but keeping its original characteristics, however as you described the Max Step or SPF need to be higher. When I do this however I end up with a sim that is very different from the original. Obviously with those two parameters set to higher values it the sim is moving faster.

Is there any other way to slow down a simulation without changing its characteristics?

Thanks.

[ivaylo.ivanov]

If you want a quick fix for this, set the low SPF to 1. However the opposite is more true - "Time scale" is preserving better the flow characteristics, and only slowing the simulation. The max step artificially restricts the flow and produces something that is more difficult to handle and understand. Time scale is not changing the flow really, the combination with the max step is producing weird result. Exactly for this reason we think to refactor the whole idea. Artificially low SPF and max step can produce faster result, however it's not very useful because you can't easily control it and get what you want. Sadly most of our example are based on using max step, and the other settings are tweaked around this idea. The good thing is that the issue is not with Phoenix itself, but more with the workflow we were using. When working with fire/smoke, the better way is to forget about the max adv step (set to a big value) and work with the other settings. You will get much more predictable results.

[dgould]

Exactly for this reason we think to refactor the whole idea.

Can I ask what new refactor/workflow you are considering? It would be great to remove the Max Step and/or SPF parameters and use some other "grid independent" parameter so that when you increase or decrease the Cell Size the sim's overall characteristics stay the same.

When working with fire/smoke, the better way is to forget about the max adv step (set to a big value) and work with the other settings. You will get much more predictable results.

I've found that the Max Step causes the sim to expand faster or slower so with a large Max Step the fire/smoke will move faster. Balancing the Max Step and SPF is the most difficult part of the software so I'm looking forward to your refactoring.

Thanks.

[ivaylo.ivanov]

Yes, this is what I mean by artificially restricting the flow. The fluid is expanding less, but not because it has such dynamics, but because it can't move more then 'n' cells per step. If you have high upper SPF limit, it will try to make more steps to compensate that. However more steps mean more diffusion, and you will start losing the small features of the flow. On other hand, if you don't have enough steps, you will end up with unnatural flow. The idea of max advection step is that if the fluid is moving with too big steps, this can mean that it's getting unstable and you must increase the SPF. So it's just another way to control the SPF in a variable way, depending on the amount of movement. This is why it's a bad idea to control both max step and SPF if you need to understand how the fluid really behaves. I think it's a bad idea to try to balance between the max step and SPF. It's really very difficult and you lose time with this. Choose one of them and make the other not a limiting factor. Fire/Smoke seems to work best if controlled from the SPF, and most of the time you can get good results with SPF of 1. Liquids can be more complicated, and max step can be more useful there.