Smoke solver scale time and output vel

[Dani F]

Hi all,

So, I've just noticed that when increasing the sim scale time (at the DOP) the output vel volume is, or it seems to be, proportionally weaker. Like if the solver would be dividing vel by the sim speed and then didn't compensate it.

I'm speeding up a sim (quite a lot) and, althought I'm not sure if that's actually a direct linear relationship the motion blur looks just fine multiplying the vel volume by the Scale Time I'm using. I haven't researched this any further yet.

But that's a kind of a funny behavior, would like to double check if this is actually something I should be expecting. Anyone knows about this?.

Thanks,

D.

Edited March 15, 2016 by Dani F

[Atom]

I wonder if it has anything to do with the fact that inside the Houdini 15 smokesolver_build2, that resides within the pyrosolver1, there is a single gas_calculate node whose Time Scale attribute is not linked to the master time scale at the top panel of the solver? So when you alter time, the sink for the smoke still remains in 1.0 time. You could test that by pasting a relative reference copy of the master Time Scale parameter on to this gas_calculate node's Time Scale parameter.

 

Also, if you plan on animating the Time Scale you might need to make sure you change Use Default to Set Always.

Edited March 15, 2016 by Atom

[jkunz07]

That's just doing a minimum operation so changing the timescale won't have any effect

[anim]

...

So, I've just noticed that when increasing the sim scale time (at the DOP) the output vel volume is, or it seems to be, proportionally weaker. Like if the solver would be dividing vel by the sim speed and then didn't compensate it.

I'm speeding up a sim (quite a lot) and, althought I'm not sure if that's actually a direct linear relationship the motion blur looks just fine multiplying the vel volume by the Scale Time I'm using. I haven't researched this any further yet.

....

that's correct, to get motion blur you expect, you need to multiply vel by your timestep you were using

 

that's because vel is not scaled during simulation at all, it's always kept in units/s for real time (timescale 1), which makes sense

[Dani F]

Thanks guys.

that's correct, to get motion blur you expect, you need to multiply vel by your timestep you were using

 

that's because vel is not scaled during simulation at all, it's always kept in units/s for real time (timescale 1), which makes sense

Indeed. I was thinking it would be related to time scale, but makes more sense that it is to the timestep instead. A good thing to be aware of in any case.

Thank you so much.