rickh

I had tried to get exactly this effect a while ago and failed. I never thought of just using a spherical combiner with Global Axis set to make the water surface dissappear outside the bucket. It really is a brilliant solution. It does its job perfectly. Just some ideas that might help. I haven't tested these yet, so they don't necessarily improve your model. - To stop the water surface appearing out the bottom of the bucket (low water + high bucket angle), just add an extra gradient combiner to your WaterMask. Set Y Start to 0.2 cm. Set End Y to 0.1 cm. Leave default Attribute 1. Set Attribute2 the same as the Spherical Combiners Atribute 2. - To make water, start off with 100% transparency, some specularity, perhaps some reflectance, and most importantly some Index of Refraction (say 1.1). Once you have refraction, transparency can be 100% because you see the water by the way it distorts light. It can look stunning. - The Ripples. Ripples in a glass or bucket often end up being standing waves which means that if the water surface is made out of concentric rings, just making alternate rings move up and down would probably look great. By this, I mean that when odd ring numbers move up, even ring numbers move down by the same amount. I was thinking of something like using expressions to move the rings sinusoidally (constant frequency) with an amplitude proportional to the absolute angle of your bone 2 from the vertical. Also leaving a very slight movement in the surface when at rest would make the water look more interesting. - I noticed that the fit of your Spherical combiner to the bucket was not great (I know it is just a test). Here is an easy way to get a perfect fit: Start a new model and make a sphere, say, 50 cm in radius and with lathing set to 12. Add the bucket and the sphere to a choreography. both should be exactly at the origin. Distort the sphere using the X, Y and Z scale and the Y translate for best fit. Make sure that sphere just touches the inside of the bucket top and bottom. Make sure the sphere doesn't penetrate outer walls of bucket in the middle. Now just use figures from the sphere directly in the Spherical combiner: Ring 1 size = Diameter of Sphere Model (in this case 100cm) Ring 2 size = Some bigger number, say 1000 cm Y Translate = sphere's Y translate in the Choreography X, Y and Z Scale = sphere's X, Y and Z Scale in the Choreography Done. The spherical combiner and the bucket will now be perfect fits. Thanks again for your fabulous idea for the water. I am looking forward to making my own bucket project work now (finally). Richard.

Korken, The regular tile pattern makes it impossible to tell the difference between reflections and detail visible through the balls. I cannot see the edges of the far balls through the leading ball either. A more diverse background and floor would make the image visually more interesting. A viewer would see the whole background directly, along with the distorted background that the balls are capturing. It would give more clues for the eye to make sense of the picture. As a matter of interest, have you tried setting transparency very high (even at 100%) and setting refraction to, say, 1.2? Once you have refraction combined with reflection and specularity, it is not so important that the glass itself is visible. Richard.

Probably not. A:M uses Single precision floating point, which means it is not great at modelling things at a huge scale. Normally, single precision will not cause an error in the 4th signifigant digit (5th or 6th digit is more likely), but it is quite possible. More importantly, if you want to animate, say, some climbers up near the peak, it is very important that you move the whole mountain so that the region where the climbers are are is close to the A:M origin. If you wanted a scene where the climbers falls off a gigantic cliff, move the mountain up rather then the climbers down (so the climber still stay near the origin. Basically, due to the effect of rounding errors with the single precision, you wouldn't want characters you are animating to ever get more than a couple of kilometers from the origin - the closer the better. Should Hash have used Double precision which would allow to animate (in real-size) a trip from the Earth to the Moon without rounding error problems? They could, but it would almost double the RAM requirements and possibly halve the rendering speed. With the current hardware, Single precision is an excellent compromise. Richard.