海面上的水花的特效,水花是在浪尖处发射出来的,目前还没有让发射参数自动化的方法,都是表达式。一旦改变浪的性质,发射的粒子就要重新写动态了。
Splash on the ocean emitted from the crisp faces on top of the wave, the whitecap, all controlled with expressions. I don't have an automated emission parameter function, once the type of the wave changes, the expression should be re-write.
第二个其实开始是要模拟一下2012大母舰被冲上陆地的镜头,特写。不过这个想法因为时间关系就停在了开始,发现几个问题:
Here's some issues:
1.Maya粒子模拟,即使是用无网格流体SPH来计算,可能数量大一点没关系,不会崩溃。可是Mesh的过程就杯具了,只要分辨率和粒子数量达到一定的高度,Maya果断crash! Maya can handle large amount of SPH particle simulation, though slow, but the mesh process is a disaster, once the resolution or particle counts goes up to a certain amount, maya crashes immediately.
2.Mesh的机制很不好,有resolution限制,这个限制会导致即使在Maya不崩溃的情况下也会出现模型抖动的情况。Maya has some issues with the mesh tool, resolution restriction will cause mesh jittering all along.
3.构造历史过多会导致问题更加严重,在流体模拟这种课题上,如果没有历史的帮助,整个模拟的效率会大打折扣,而且有些效果成为不可能。Construction history cause problems when there are a lot of them, but they are important for fluid sim, some effects will be impossible to do.
4.Maya的SPH计算的结果并不好,压力不够,张力不够,能量传递有时候有问题,像我同事所说的,可能Maya粒子流体只适合模拟大鼻涕一样的东西。SPH in Maya doesn't work well enough, problematic pressure, tension, energy conservation, transferring.
5.再次希望,Maya能够推出多质量的网格流体模拟。现在的流体都是气态的,或者模拟两种质量相近的流体。而气体/液体这种,才是我们迫切需要的,毕竟大规模流体我们不能用SPH。(Houdini 11有了自己的FLIP Solver,Realflow 5有了Hybrido混合解算器,Maya估计也快了吧) Hope that maya will have multiple mass grid fluid simulation. For now all simulation are gaseous, all 2 same mass kinds of fluids. Gas/Fluid is what we need the most, we can't use SPH for large scale sims after all.