Everyone at this point is subsidized by general tax revenue; neither transit nor roads pay for themselves, so I'm not sure it makes sense to say one of them is subsidizing the other one. It may be true according to some ways of accounting that transit is more heavily subsidized than driving, depending on the specific transit mode, but if you actually charged rail users what it costs to run a rail system, many fewer people (as you say) would ride the train, and they just wouldn't fit on the roads in major dense metros because cars are so space-inefficient, so you'd need to build way more roads, probably seizing currently-not-road space by eminent domain, and you'd want to account for the loss of economic activity that came out of tearing down a bunch of buildings to replace them with highway. All of which is to say, I'm skeptical that an actual user-pays model would save drivers money over transit riders, at least in dense, transit-heavy cities.
Another way to think about it: you can view drivers paying for transit as a subsidy, but you can also view it as paying for decreased traffic, from which they directly benefit.