Maya Fluid Effects Dvd now online

Hi Duncan..

Thanks! I think I should probably post your reply in some forum.. Very informative! :)

Substeps can affect the simulation in a couple of ways.

- by using smaller time steps the flow can better resolve incompressibility and high velocity flow, which can sometimes dramatically change the simulation by reducing artifacts (Maya fluid is designed to work with large time steps, but instead of blowing up it tends to have excessive diffusion for large steps)

- In some situations small changes to the simulation can have a dramatic effect on outcomes (the butterfly effect)

- emission currently doesn't occur on substeps, which can lead to some differences, although total emitted should stay the same. However emission does occur on subframes when caching with oversampling.

We try to ensure that the behavior of the fluid is roughly the same for different timestep values, but there tends to be a difference because we have a rather large default timestep.(one step per frame) However there may still be some behaviors that are not properly adjusting for the timestep(i.e. bugs)

The quality attribute controls the amount we iterate on the Poisson solve, thus affects the incompressibility of the flow. Low quality values are less different when using the high detail solve method, because it uses the pressure from the previous timestep as a starting value, while the default method always starts from zero pressure.

Thanks :)

There aren't that many jobs in the academy for high-energy physicists.. You have to be some kind of genius to get one..
What's still confusing me about the simulation engines for the fluids and nparticles is that some physical attributes are controlled by parameters of the solvers, like substeps and quality.. Their affect on the behavior isn't something I could have guessed..

And thanks again!

Nizan.

The Navier Stokes model is a 2d incompressible flow, not a heightfield water model, and thus will not yield natural looking wakes. Sometimes one can get nice looking results with the ns model for a height field, however. In some cases it can create the look of standing waves in a flowing river, for example. Thus one can get things like the buildup of flow around bridge pylons ). The reynolds number is determined by the viscosity. (we use a formula to convert the viscosity into a reynolds value internally for the ns model, but I don't think you can really calibrate it very easily to water and scene scale. It is best just to play with the values to get a good looking effect, not try to enter emperical values ). Also sometimes when using the ns model for a heightfield it is useful to relax the incompressibility of the flow by lowering the quality attribute. Forward advection may also be useful. One can also get interesting effects by using negative density gradient force. (which is almost like a shallow water model, but very crude)

I think a physics PHD is a great asset when getting work as a TD in simulation or research at one of the larger animation houses that support this kind of work, or in a games or software development company like Autodesk. However getting work in the 3d market( at least in computer animation ) is never easy, and jobs have been a bit tight lately.

Cool, thanks :)
And another question (let me know if I nag too much). I was trying to use the wake fluid to create "swelling" near an obstacle.. It seems the spring mesh solver gives too-linear results, so I switched to Navier-Stocks.. But I'm having trouble figuring out the scales.. How would I go about calculating the Reynolds number from the flow? Do you happen to know if that swelling is a result of depth-waves hitting the obstacle or just something related to the viscosity/compressibility of the flow?

Btw, do you think it'll be easy for a physics phd to find a job on the 3d market? :)

Thanks,

Nizan.

I don't think the ratePP is what you want... it controls the emission from particles when particles are the source of emission.
I'm not sure about your shading problems, but a suggestion would be to not use the ocean shader but instead do everything using the ocean texture as displacement and bump, as well as opacity mapping for a layered foam shader( using the outFoam from the texture).
The ocean shader has its own built in bump and displacement as well as color mapping, which can make it efficient, but it can be somewhat non-standard to use in shading networks( it can work well but you need to understand it ). One thing... if you are raytracing then make the environment ramp on the ocean black because it will add on top of raytraced reflections, washing out shadows and reflections.

Hi Duncan,

Cool idea.. I will give it a try! I was also thinking of controlling the ratePP with some volume field's inputForce (where the volume is set to the peak's height), so I can set the rate directly.. Will it work? Never tried working with ratePP but it seems to create a strange table in the attribute editor.

And another unrelated question - I was rendering an ocean with some objects which were assigned a mia_material, using mr sun&sky, and there are a few strange things - I can't get specular and shadow maps (maybe not ocean related, still need to figure it out), and the AO map only applied to the mia objects.. The ocean appeared flat.. Do I have to bake the displacement? I was trying to assign a mia to the ocean surface and use the displacement map from the texture as a standard_bump, but then Maya crushes.. Is that a bug?

Thanks! :)

Nizan

I think your problems likely not the "if" condition, but rather thar we need to emit a very large number of particles, most of which are then deleted because they are below the emission height.
I don't know if the following would be any faster, but you could try it:
In maya2011 the ocean can be converted to poly using "convert displacement to poly with history". Then do a Boolean with a plane so only the wave peaks are left and emit from that.

Hi Duncan..

Not sure if you'll get this, but here it goes.. When working with emitting particles for spray, it seems the if's and setting the lifespanPP slows everything down considerable.. Is there a way around this? (i.e. - without using an if statement)..

Thanks!

Nizan.

Hi Duncan..

Not sure if you'll get this, but here it goes.. When working with emitting particles for spray, it seems the if's and setting the lifespanPP slows everything down considerable.. Is there a way around this? (i.e. - without using an if statement)..

Thanks!

Nizan.

Very helpful....
but i have a question...how to create a sea tide effects washing the beach???

Hi Duncan - a quick question: is there good training material available for volumetric particle rendering? using fluids as shader looks really nice, but takes too much time to render (as soon as those "spheres" intersect).
I want to take full advantage of the particle cloud shader.

I am getting the same message as Jesper, and I have not upgraded to Maya 2010 yet. Is there a problem with the training section?

Iam a subscribed member but I cant find any dvd at all.

"There are no courses associated with your product(s)."

can that be because I upgraded to Maya 2010?

As I noted in an earlier comment, Autodesk no longer sells it. You could try to find it someplace like ebay. Googling "Maya Techniques | Maya Fluid Effects" brings up a few hits, although I can't attest to any of these sources.