It works great. The only real problem (and it is hella annoying) is that the sound overrides the mixer volume for the period that the voice nav is talking. It actually sorta does this in Windows, but it doesn't come out as bad.
I suppose if I were to look for a /dev/dspX multi-mixer, I could fix this problem, but I'm not sure where to look. 'dsproxy' is in the right direction, but I haven't taken the time to properly work it out.
Something like this ought to exist, because it would handle the long-term problem of racing /dev/dsp access between, artsd, esound, and so on.