She puts a lot of ds and bs where I'd put ts and ps (and she reverses d and b a lot too, just to add to the fun!) because I guess she hears them as voiced when they're between vowels. I don't, and I don't think I say them that way either, but she does.

Especially for d where it "should" be t, that's not that surprising - I think a "flapped", voiced pronunciation of /t/ between vowels is common in American English. (Though I'm not sure whether that makes "ladder" and "latter" homophones, or whether there's still a difference.)

See http://en.wikipedia.org/wiki/Intervocalic_alveolar_flapping and http://en.wikipedia.org/wiki/Withgott_Effect, for example.