Is it possible to capture all/any audio played by a PC into a system.io.stream, so that it can then be run through speech recognition (System.Speech.Recognition.SpeechRecognitionEngine)?
Essentially I'm looking to pefrom speech recognition on any audio on the client PC, google seems to suggest that capturing a stream like this can be done using Microsoft.DirectX.DirectSound, however I cannot honestly determine how. Any suggestions would be greatly appreciated.