multiple levels of infill

I am sure it has been discussed before but I am not getting hits.
Anyone have any suggestions for printing with different amounts of infill in a print. One way I thought of which would get you a bit more flexibility by sectioning off larger areas of infill in cad in such a way that they would be sliced as solid because of their size. This however would only give you solid and your given sliced infill % not more options in terms of different %, it could be very time consuming dividing up the areas of the part so it prints solid.
It just seems like such a waste of fillament printing parts which you need a certain strength in places but you have to have the strength everywhere. Fillaments expensive on larger parts!
Would it not be possible to add a feature to slic3r that you could select parts of a model and assign infill and number of walls? It seems like quite a hole without this.

Edited 2 time(s). Last edit at 11/05/2013 06:02PM by dave584.

i always though infill should be more like bone, the more dense the closer to the outside you get.

Quote
dissidence
i always though infill should be more like bone, the more dense the closer to the outside you get.

I agree that could be a good option. Also support material is very wasteful as its only there to provide a structure for overhang to sit on. you could have a very sparce structure and a few layers before getting to your model the support could begin a bridge with a tighter pattern and then another bridge the next layer with a tighter pattern so you have a flat to support without using nearly as much material. Its not going to make much difference for parts say 30mm x 30mm but for 100mm or more you could do really big gaps to begin the structure then make them tighter at the top.

Edited 1 time(s). Last edit at 11/06/2013 06:28AM by dave584.

Ah, the variable density infill! A group from the MIT Media Lab was working on that, but it seems like either they didn't publish further results or went on to do research on other stuff.
However, I am planning to do my bachelor thesis on variable density infill, too, so I have already made some promising tests. Actually, when using skeinforge, slicing the variable density models is way faster than the normal infill method because you only generate perimeters.
The trick so far is to use blender's cell fracture plugin to divide the object up into individual cells and subtract them from the original object. You can also set the distance between the cells in order to generate a multiple-walled infill for example.

Thanks for the info uGen. I have not used blender and its free, will be giving it a go.