Okay, first post here, been a lurker for a bit, but decided to come out of the wood work to post an idea.
Voice recognition whilst playing music is a hard one, as the computer will listen to all the sound it can hear, music + your voice, i have an idea that may help this, but i'm not sure, as i've not done alot of work with speech recognition!
Basically, your computer is controlling your speakers, so it knows what sound should be in the car at any one point. Now, it constantantly listens to this, and compares the sound it hears, to the sound that its playing, and takes away the sound its playing from the sound it hears, therefore your are left with the sound that should not be in the car. This you perform the speech recognition on.
Is this possible?
Appologies if i just wasted 2 mins of your time!!
edit: another quick question.
Is there a GPS software with an SDK so i can ask it if i'm within 100m of a petrol station, if so, dont shutdown on next engine off?