2
$\begingroup$

just working on this extremely fast moving smoke simulation, which is a trail of smoke emitting from a cylinder constrained to the back of a drop pod flying towards a planet.

And, the smoke comes out in puffy stuttery line like this, enter image description here

which is not a continuous stream of smoke. How would I get the smoke simulation to be a consistent trail behind the drop pod?

Thanks!

$\endgroup$
1
  • $\begingroup$ Apart from the things in my comments under the answer which have to be considered, maybe it is best to upload the file (reduced to the parts relevant for the smoke simulation) here: blend-exchange.com and edit it into your question to inspect it further. However I have no more time to check it today, have to do that tomorrow. $\endgroup$ Commented Mar 5 at 15:40

2 Answers 2

3
$\begingroup$

in the domain, increase the minimum time steps.

enter image description here

$\endgroup$
1
  • 1
    $\begingroup$ Setting a minimum of 10 is unnecessary when the maximum is 4, the tmesteps will not exceed the maximum. So if the cube was smaller (for example so that it fits 5x into the distance it travels between two frames) then there would still be gaps and they will remain no matter if you increase the minimum to 20 or 50... $\endgroup$ Commented Mar 5 at 13:32
2
$\begingroup$

Instead of increasing the Timesteps > Minimum in the Domain, I would increase the Sampling Substeps on the Flow object. The reason is, even with less computational effort (i.e. shorter bake time) you'll get better results. This is especially true when you have fast and slow moving objects within the domain - the higher amount of substeps would mostly be required for the fast objects, but the timesteps will be used to calculate all objects regardless of their velocity.

sampling substeps

Here are some comparisons, a cube as Flow object moving quickly through the Domain. First at the top, the default settings with Timesteps Minimum at 1 and Sampling Substeps 0. A lot of gaps between the smoke puffs.

Next the timesteps increased to 2, unsurprisingly there are now puffs of smoke in the gaps between. The result is still not good, but the bake time already increased by almost 50%.

At the bottom, instead of increasing the timesteps, I set the Sampling Substeps to 10 on the flow objects. The result looks now smooth and the bake time only increased by 10%. Even if I wasn't satisfied with the smoke yet, doubling the substeps to 20 would only increase the bake time by 20% compared to baking with default settings.

smoke trail

Nevertheless, increasing the Timesteps Minimum can be useful, too. Because even though for quick moving objects the substeps might be more important, a higher Timesteps Minimum value can result in more detailed and therefore better looking smoke (personally, I always use a minimum of at least 2 instead of the default 1). But you have to keep in mind that it increases the bake time significantly.

$\endgroup$
3
  • $\begingroup$ Hey, Thanks for your response! I've actually got 200 sampling substebs, and 15 timesteps minimum. Yet It still looks like the image that I posted. I'm actually not sure what to do, I've reset my domain multiple times, baked the MAX sampling substeps, MAX timesteps yet it continues to provide the same result. $\endgroup$
    – Falrte
    Commented Mar 5 at 14:45
  • $\begingroup$ @Falrte As I mention in my comment under Chris' answer, increasing the Timesteps Minimum only helps if you increase the Maximum accordingly, because the steps will never exceed the maximum value no matter how high you set the minimum. The other thing is: where is smoke emitted? All around the object? At the front? At the back? If it is at the front and it has to move around the object, this will have an influence too since it looks like a kind of "shock wave", which other than my example has no completely empty space between the steps. $\endgroup$ Commented Mar 5 at 15:02
  • $\begingroup$ @Falrte You can see in my first bake with timesteps 1 per frame, the distance between the smoke puffs is the distance the cube travels between frames, and since its size is a lot smaller than the gap, it will take as many substeps until it looks like a smooth flow (in my case 4 would have been enough). So how fast is your object? How large are the gaps between the emitting areas from one frame to another? If everything seems as if a smooth flow appearance should be possible, then it's probably more like a shock wave effect as mentioned in the last comment. Is there an effector object as well? $\endgroup$ Commented Mar 5 at 15:05

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .