For my carputer, I plan to run voice recognition software. Thing is though, I don't want to have the carputer picking up song lyrics and think it's a command. My solution to this is to have a toggle switch for whenever I want to tell the carputer something. Here's what I want it to do when the switch is on:
1.) Allow microphone input.
2.) Mute soundcard going to headunit (I will have a second soundcard outputting to a small speaker in the dash, to make confirmation noises for when it's successfully understands a command)
3.) Turn on LED in dash.
What would be the best way of going about this? It sounds like I *should* be able to do this without any software related.
What voice recognition software are you using? I think it would far easier and better to do this in software as I know at least the microsoft voice SDK and all software built off it supports this function directly. The software shuts off the recognition completely so that it does not consume any CPU cycles waiting to see if someone is talking. You could also just get / write some software that would mute the Mic line on the computer if you don't want to mess with the voice recogntiion software.
If you absolutely had to do this in hardware and only hardware you could just wire the mic input to a relay that was triggered by a 3 pole toggle switch so that it would be connected when you press it and unconnected when its not a second relay would be conncted in the same way to the LED. Another relay would be used to wire the sound card output to a relay wired to the other end of this toggle switch. You would have to hold the button down during the whole time you spoke, to get around this you could build a timer circuit to trigger the relays so that the mic would be connected for say 5 seconds after you pressed the button. This wouldn't be nearly as good as what the software could do though since it could switch off the mute as soon as recognition has finished.
It's voice recognition... how would you possibly do this without software? It's practically ALL software. You don't need a second sound card... just pause whatever is playing and play the confirmation tones. Don't use a button for commands, use a trigger word. The whole convenience of voice operation is kinda negated when you need to press buttons to operate it. As for the LED, wire one up to the serial port.
I'm using nitro voice flux. They way I'm using it, I can't have the confirmation word for when I want it to do a command. When I want to search for a file I have to spell it out, which basically types it on the screen and gets recognized by SneakyAmp.
I really don't see how doing it all in software would be easier, if I'd have to pause the headunit to do that anyway? Seems like the switch would be just as easy, and not be open to receiving bad commands. I want the second sound card because I want the computer to talk to me out of JUST a small speakers behind the dash. I don't want it going over my system speakers.
What year of Firebird do you have god_of_cpu? I'm putting this in my 91. I just took out the whole dash and put in one from a 4th gen 95 Camaro. The mainboard fits just about perfect in the spot for the airbag.