IMO none of this matters much if you keep the sound pressure levels (spl)within a reasonable range. There are three sources of vibrations:
1. Self induced (speakers being the obvious example). It's been said that in poorly braced speakers more sound energy can come from the cabinet than the drivers. Some designers advertise that they account and even use this fact in their design. Personally I don't want all my music colored by the cabinet, so I'm a believer of inert, massive speaker cabinets. The heavier the cabinet the greater energy will be required to create a resonance at a lower frequency. The same applies to transports and any component that "hums". Cabinets should be well braced and stiff. Every component within your piece of equipment (resistors, caps, chips, etc.) are mounted so as to be thought of structurally as cantilevers, the worst possible scenaro. Thus the interest in tube dampers, adding potting clay to chips, sheets of sound deadening materials for equipment walls, and adding weights.
2. Air borne. Unless you're talking about extreme spls or your components are shaped like an airplane wing there isn't much here to talk about. Tone arms are the exception, but hopefully you have a dust cover that will resolve the issue.
3. From the support itself. Soft wooden floors are the usual culprit. The worst ones are those without drywall or similar attached to the underside as this adds stiffness and the blocks resonate vibrations from below. Some have reported 20 dB peaks at certain frequencies caused by wooden floors over crawl spaces. Raised platforms are perhaps the worst example.
Resonance is another key and has three factors of it's own: mass, material, and shape. By shape I include span, thickness, and bracing. Obviously metal rings like a bell. Adding a different material to create a composite is one option discussed below. And mass should be easy to understand as what we're discussing here is transfer of mechanical energy.
Another point to be made here to think in terms of resonance frequencies. The theoretical model is a series of attached springs. Some are big and stiff, others small. The massive, inert ones would only vibrate at very low frequencies and under heavy loading. Even the Earth is a spring (and has a resonate frequency of 1 - 2 Hz). This is how composites work as one material can filter out at least some of the energy at a given frequency from reaching the other material that has a different resonate frequency.
Back to your speakers.
A perfect coupling (100% energy transfer at all frequencies) between speakers and table would create a composite object with greater mass (a good thing on two counts), but the table is designed like a wing (a minor bad thing), and probably isn't as inert as the speaker itself (another bad thing). Bolting the speaker directly to the table could create this perfect coupling. Note that nearly any product available relies on gravity for the coupling (a relatively weak force).
At the other end of the spectrum would be to float the component (an AC vendor offers such a device that uses magnets). This eliminates the support induced vibration source, but leaves the component on it's own for self-induced or air borne vibrations. Of course putting it inside an air tight cover and having remote control would take care of the air borne vibrations, which might work well for DACs, pre-amps, and power amps.
I don't care for roller supports as they do almost nothing in the horizontal plane (and they're a PITA). These devices, and spikes, create a composite with the component and the support so the material used does affect how vibrations are filtered (honestly though I can't understand how the resonant frequency of different metals in these shapes would make an appreciable difference). And I can't imagine how the shape of a spike or roller cup would make a difference either. Note that a spike disc or roller cup adds another coupling interface that most fail to consider.
The soft coupling devices work in three dimensions and don't rely on ringing metal. Not sure if Gary still sells Panda Feet. I like his stuff and his site helps to explain some of this (they're inexpensive sorbothane feet designed for given load ranges). They're the only such device I ever bought. But we both agree that the Panda Feet seem to have little benefit for digital equipment (DACs and amps). The Vibrapods/Vibracones avoid the expensive snake-oil route and now offer a variety of options.
OTOH I've never heard a difference with any isolation product, so much of this seems to be only theory to me, although I try to be open to possibilities beyond what generally accepted science can explain. (For instance I can't figure out how a brass spike is sized to different loads.)
