Reducing 'splashiness' on sim / exotic parameters

[o.delellis@gmail.com]

Hi,

where do i start...

i am working on 5 shots, and i have a sweet spot for the look of the fluids, ATM they look pretty real, with the right amount of splashes for this animal swimming on the water. good.

however there is a pretty big gap between the object and the fluid, i tried using negative values on 'surface offset' on the objects,but this is not working, so i am forced to reduce the cell size.

so sims are slower, will use a TB per shot iteration...fine, i can live with that.

The worry is, in RnD i did earlier in the project, the bigger the cell size, the more subdued the fluid movement is, so the smaller the cell size the fluid tends to splash more.

i am looking at a preview of the few frames simmed so far and comparing it with the 'good' sim i did earlier and seems that there is more water splashiness.

so having little time to do another bunch of RnD i am asking if any of you have a go at comparing:

HyEmitter:
surface sampling (maybe having a value of 1 rather than the default 2), will it make the fluid less splashy? or does it not matter as the cell size is smaller?

HyDomain:
min / max sampling, have you used this with the same value to control /reduce the particle number per cell?

what i am trying to figure out is if i can have a similar number of particles with the new cell size (0.08) than what i had with the old cell size (0.13) by tweaking the parameters above, would the simulation look / behavior will be similar?

i will be testing this, but was looking at saving some time or get my confidence levels a bit higher.

the other option would have been to use a good old Drag Daemon, but it doesn't work with H2.....

[LuisMiguel]

Hi,

The gap can be due to you have different values of cell size for Hybrido and object. Both cell size values should be the same if you want to adjust the fluid to the geometry. Decreasing the value for the cell size for hybrido will increase the number of particles as well, but the cell size for the geometry should be decreased as well. The best way to do that is modifying the Cell size global scale value, so it is a multiplier for all cell size parameters at the same time.

I advice you to leave the Surface sampling parameter to 2 for the Hybrido Emitter because the fluid will have more detail in the surface adding more particles there.

About the hybrido min/max sampling it is not for reducing the number of particles. Imagine you have a simulation with fluid and objects interacting against it. In the simulation is possible to have air cells, it means no particles, in the middle of fluid, but it would bad depending on the situation. So, the min/max will add particles for air cell when the fluid situation is required, and the particles will be added with a minimum of 8 and maximum of 16 by default.

luisM.

[o.delellis@gmail.com]

hi,

i have been using the same cell size for the domain and for all the objects.

i have increased the FPS, i completely forgot it was there...im hoping that will slow down the particles a bit.

doing a preview now,,

would you know when would there a patch for the Drag Daemon that works with H2?

[o.delellis@gmail.com]

okay,

simed a few hundred frames at 125 fps, the old sim which was a bit too fast (was simed at 100fps), so this new one at 125fps had the desired result of slowing down the particles, a bit too much perhaps, but the issue now is the particles are a lot more splashier (see pic)

fps compare.jpg

no other parameter has changed.

i could sim at 112 fps to deal with the speed of the fluid, but i would still get more splashes that the 100fps version has (which already is too splashy) and i can't see a way to decrease that.

so the dilemma is that i need a smaller cell size to have a smaller gap on the particle/object boundary, but that changes my particle splash ammount and the speed the particles react.

i am surely missing drag force and a compressibility parameter.

i am setting a 75fps sim as a last attempt before going SPH but i predict particles will move faster (bad), with less splashing(good)

[o.delellis@gmail.com]

ooooooook,

run some more tests and the negative values on surface offset DO seem to be working, careful with cranking them too much.

(as something to investigate further, i wonder how Rf deals with objects 'solid inside' when they are not closed meshes, as a creature with a mouth/ eye holes in the geometry)

one thing i was doing, maybe H2 didn't like, is to link a bunch of SD objects as a group to the domain in the Relationship Editor, im too lazy

Also, re-read the manual a bit closer and found out Time Scale in Simulation Options...yes!!!!!!!!!!!!!!!!!!!!!!!! exactly what i was after in slowing down/ speeding up a sim with out using FPS.

i will finish the 3/4 of the wine bottle and going to bed, confidence a bit higher

.... zzzzzzzzzzzzzzzzzzz

[o.delellis@gmail.com]

oh, one more thing...


it might be a good idea to be able to control the distance on 'shell' mode for objects, from cell size to sub cell size...it could come in handy as sometimes it doesn't matter if there is particles inside your collision object. could be a 'brute force' sort of collision