3 problems, solve them in order.
First, can you write a windows app? If you can program in a language Visual Studio likes, you are set, but otherwise mixing audio buffers and networks is a tough start.
Capturing audio. Look at the SDK example Multimedia\audio\CaptureSharedEventDriven and see how it works. It captures audio and writes it to a file, but it will get you looking at the right header files for audio work. Then you need to work out your buffer; circular array should work best for this. More of the audio API is available http://msdn.microsoft.com/en-us/library ... 02(v=vs.85
).aspx and you might look at the audio programing guide http://msdn.microsoft.com/en-us/library ... 99(v=vs.85
).aspx and focus on both the endpoint (microphone and speakers at the other end) and the stream management (how you read from the mic or output to speakers).
Network stuff is last, and I'm not the person to ask about that. I don't know what the stream protocols look like, or how to encode to them. An MP3 stream should work, I would suggest google for something like a streaming audio library. That way you have a standard output. Or you might create your own protocol so only your own device and app could listen at the other end; though security says that's a bad idea it might be an interesting project. Piping raw PCM WAV across your network might work, but you'll need a large buffer and have to account for odd latency issues if you have any network problems; while compressing your audio buffer into an MP3 or other format stream will require the above work.