[Water Simulation Plugin]

bump maps?

 

no. bump/disp maps wouldnt help much to rid the gridded look, only manipulating gammas and increasing mesh density could achieve that.

With all due respect, I believe both bump and displacement maps will help to eliminate the "grided look".

• Firstly, since the cp:s only have to hold the low frequency data, you can keep the mesh density down.
   
  
• Secondly, since displacement maps subdivide each patch it allows for even lower mesh density with retained quality
   
  
• Thirdly, bump maps have as good resolution as the map itself. This map, holding the high frequency data, could add to the natural look with subtle changes in the water surface. This would help to hide the "grided look", should there be any left.
  

In order to have high resolution maps you also have to simulate for that higher resolution, however. This shouldn't be too big a problem in 2D though.

i also have no plans of doing that, even though i think it would be possible with the sdk.

If that is indeed possible with the present SDK I would have considered it at least. If not only for the suggested bump and displacement maps, but also for color maps that could be white at high gradient changes for instance

 

i'm using a fast 2d solver which i have developed myself. it actually does just a few things:

1. every cp has a speed

2. every cp has a height

3. this speed is modified by surrounding cp's heights

4. speed is dampened

this offers enough rooms for parameters (dampening, height scale and the reaction part) while staying  very simple.

This sounds like a finite difference solver for the 2D wave equation. This yields a decent approximation for infinite depth water, but I think it should be possible to approximate topography by depth dependent damping factors for each cell.

i also have talked to martin about how to bring a:m water sim some steps further and i'm going to do some research about true water simulation and possibly implement a generic water simulation plugin that would simulate any kind of liquid in any kind of situation.

That sounds promising. Do you have any ideas, as of now, on what method such a generic fluid solver would be based on? Since you have been in contact with Martin on this subject, maybe you have already discussed such details. I'm just curious, you know

/Anders

[Water Simulation Plugin]

the plugin currently just moves the cps.

[...]

yes the gridded look worries me too

[...]

Marcel, do you think it would be possible to, in addition to moving cp:s, also create a bump (or even displacement) map movie. I think that would get rid of the gridded look.

The procedure would be something like: Low frequency data for moving cp:s, Medium frequency data for displacement maps and high frequency data for bump maps.

 

I've not studied the SDK enough to know if this is even possible, but I guess you have and since you haven't used that approach it probably isn't.

 

While on the subject, what is your approach for simulating water in this plug-in? Is it a simplified solver for the two-dimensional wave equations or maybe even the shallow-water equations? If so, are you planning to include fluid parameters such as density and viscosity? Also, are you planing to let the bottom topography have any influence on the fluid movement.

 

Of course I realise these "improvements" will not be available in the first version. I'm only curious of what future plans you might have for this much longed-for plug-in.

 

/Anders