See
http://uhfmag.com/Issue63/soundproofing.htmlIt explains a lot, in technical, but understandable language.
An important concept is that you do not want to transfer sound energy to the floor, because it will travel (just like the wave the energy is in the form of) as a wave along the floor.
The more contact area between any two objects, the more sound energy is transferred between them. If you (theoretically) had an object you could use to separate the speaker from a floor, such that the contact point was only one atom (or even a molecule) in size, then any energy transferred would have to do so serially through that tiny contact point. Needless to say, it could take decades for the full energy of even a single low frequency sound wave to be passed through that point.
Of course, you could not have such a tiny point that could still serve to support the speaker, so you have to settle for the size of the contact point on a cone or a spike.
Now, think of what happens when you use anything the size of the speaker base to connect to the floor. if the waves are longer than the thickness of that something, you will couple the speaker to the floor for their full amplitude (but if shorter, you have a chance at absorption,depending upon the material).
The same is true of spikes or cones, but the contact area on one end is so small, that so little of the wave's energy can pass through that contact area and so the amplitude of the coupled wave will be virtually nil.
Think of separation, not absorption, for low frequencies. Effectively absorbing low frequencies could require separation objects that are too massive to be practical.