From work I have done with modal analysis and vibration analysis of large vibrating machines, I imagined (reasoned) that different combinations of speaker isolation method and floor construction might give very different results, some of which could perhaps be audible, but most would definitely be measurable.
A quick insight: these vibrating machines weigh 60,000 kg and vibrate in the vertical plane at about 6G at frequencies between 20 and 30Hz. To help the supporting structures survive the fatigue loads, and reduce the destructive vibration in neighbouring equipment we use isolation frames to mount the machine onto, which then mount to the structure. Both machine and frame are supported on springs and by tuning the mass of the isolation frame almost perfect isolation is achieved at running speed. In other words the structure doesn't move while 60 tonnes tries to destroy itself at 6G & 1,500rpm. Of course this is a different mechanism to a speaker because mostly the running speed is constant, but it has made me wonder about testing the same concept/ idea with my speakers.
I have been experimenting for a while with isolation versus direct coupling, using various simple isolation methods.
My premise was that sound is generated by moving air, and by inference also indirectly by objects vibrating which are exposed to the same air. If you feel the floor or furniture moving it is creating sound and or noise and or distortion - and naturally the question would be is it loud enough to be audible. But first I just wanted to see if I could measure it.
So to try and measure this effect I used REW with sweeps, pink noise, and also played musical content in the nearfield on the same speaker with different types of support. Both inside and outside the apartment.
I also recorded musical content with big dynamics (lots of woofer excursion) using a usb mic in nearfield, and then compared the result against the original file using DeltaWave - with the different types of speaker supports. The most obvious results were in distortion (when something vibrates during a sweep it shows up as distortion in the measurement), and also shows in the peak levels of transients (spectrogram plot in DeltaWave). Anecdotally, both the floor and the entire speaker cabinet vibrate much much less as the same SPL when isolated on Sorbothane (who knew...). I wonder if this vibration is actually audible, or it just manifests as reduced clarity, but for now it exists, and it can be measured.
Anyway, as a result of all this I have been using four 2" Sorbothane half domes under each speaker - with the mass of the Revels this results in about 30% compression on the front pair, and about 50% compression on the rear pair. Meaning they are within their nominal range of function as I understand it. There are pockets of information available on this topic, and certainly from first principles it seems you would want to isolate the vertical direction far more than the horizontal. So I am looking into the next evolution of isolation to try.
If work, life, and listening time permits I will try and pull together a post summarising what I managed to measure - with actual measurements and stuff.
