Vandersteen 5a battery biased crossovers

I am not sure what caps are used but any high voltage signal path application for caps generally requires high quality caps.

Most speakers use high quality polypropylene caps for the crossover ($5 to $10) as do Vandersteens (I think) - these do not need reforming like electrolytic caps (although like any cap they drift with use with most drift occuring in the first 100 hours). So I am not sure what the battery actually does unless? As for peformance - within a few minutes the capacitors within your speakers should perform optimally so I am not sure the advantage there either. (The rubber surround on the speaker drivers will also ease up after a few minutes each time you power up too and thermal heating of the voice coil will change the way the speaker behaves too - do you have a "block heater" on the driver motor assembly as we Canadians have on our cars? It would be a sensible addition if one is worried about capacitor warm up)

I have thought a bit more about it and since a battery typically has about 1.7 ohms of internal resistance and since it must be part of the circuit somehow then you may need to worry about it. A leaky or corroded battery or a change in the battery internal resistance over time might affect the sound. If you think about it - a chemical battery will likely change properties over time quite a bit more than a stable polypropylene capacitor - so it might be advisable to replace/check this regularly....

I did some digging around. JBL use this Vandersteen idea in their K2 speaker. As you are probably aware capacitors vary in quality and linearity. If you bias a polypropylene cap then the distortion products (due to non-linearity) will still be there but they will all be even harmonics (more benign to our ears). Also the circuit should be such that it won't matter about the battery aging (normally connected to a 1 K or more resistor). It does mean that the circuit requires two much larger (more expensive) capacitors in series of twice the value of what you would need if you used one capacitor. Since capacitor non-linearities tend to increase with size, a lot will depend on a the careful choice of components if it is to be an improvement over a single capacitor.

It also opens the door to using polarized capacitors in a crossover although I don't think Vandersteen does this (he uses polypropylene)

Anyway - interesting stuff - I hope this helps. I have found an AES paper on this subject if you are interested.