Given one, two, or three square images, produce a 3-dimensional rendering of them as a block in glorious SVG.
The svg has two layers. They contain identical shapes, but the lower one is slightly blurred. Although the shapes are aligned pretty well, some rendering engines create a 1px gap between shapes due to rounding. The blurred object helps fill these in. A mask is used to make sure the blur does not cause a halo around the edges. The seams still aren't perfect but they're pretty good, man.
Because each pixel of each face produces two shapes, you should probably stick to small source images.
