river

[drosselmir]

Thanks for the update!

is it using smooth conservation and classic advection too?

[Ivaylo Katev]

yes, smooth/classic.

[Da_elf]

question:
if the river was not straight would there be any way to have multiple grids that can more efficiently encompass where the river flows or will you have to setup one huge grid for the entire area and what kind of effect would this have on memory and sim time. or even with a river with bends that comes to a waterfall then the water drops 10 feet into a pool then continues on.

[Ivaylo Katev]

no effective connection yet. you can fake connect them by the cascade technique, but the particles will be not transferred

[drosselmir]

yes, smooth/classic.

OK. So after some further experimenting, I'm getting much better results in my simulation with a min/max spf of 6/6. Max step is at 8. and I have about 30mill cells.

when I used a fixed spf of 2, I was getting unnatural results with the crazy accelerating waves.

I'm trying to find a logic in how to determine spf with maybe... how fast my water is moving?

any advice?

[Ivaylo Katev]

yes, if the spf is too low, you will get these waves. the optimal spf is the most important decision for the liquid simulations, no exact formula for it, but the first argument is the simulation scale and the liquid velocity. the bigger is the scale, the lower is the spf. the bigger is the velocity, the bigger is the spf.

[drosselmir]

I love this info. Thank you! I am definitely starting to feel better about the liquid simulation in this respect. My next mystery is the foam behavior. I will soon do some simple scene tests, but I'm having an issue with keeping the foam floating on the water surface. Mine always seems to want ultimately sink. My current simulation has over 10 million foam particles, yet I can barely seem any of them because 95% are under the water. FYI it is a river with dirty water, so its more important that I see them floating and moving on the surface than swirling under the water.

I have my rising limit set to .5 and falling limit set to 50 like the "pool" sample scene. should I just set the falling limit to a very low setting?

Thanks again.

Dave

[TSR_trix]

I would try first with default values, which in many cases do the job but if its not enough try to increase rising limit. For example rising 100 falling 1000 wont let any rebels to stay under water in my test scene

[Ivaylo Katev]

the most common problem with the underwater foam is not related to the simulation, it's a question of rendering. you just can't keep all of your foam above the water! yes, the rising limit can help, it allows the foam to move upward faster and to reach the surface in less time, but when the liquid is moving rapidly the bubbles always will alternate their position above/below the water. so, phoenix foam shader has certain helper options that are made with this in mind. first, you have to set in the liquid slot the simulator, this is very important and is suggested automatically by the software. however, this may be not enough, i can't help more without a shot, but in some extreme cases you have to render separately and to compose.

[bos]

Any update on the tutorial for this?

[TSR_trix]

Im sure you can achieve similar results by using either the initial fill up or a brush mode object. A source or a two will make the flow movement to the river.
What went wrong with your tests?

[bos]

I think it was the scale. I didn't see the post on the 10m scale I have updated the sim with a larger scale. Using two sources now at the top of the river. Getting a little better results. I'll see how this goes for now

[TSR_trix]

Ok good. Remember to use enough big values with SPF if your resolution is over 10 million+.