Implement speech-to-text recognize audio api using websockets

[TerryTsai]

[RMichaelPickering]

This is needed to bring the UE plugin capabilities in line with what is available in the IBM Watson SDK for Unity. I would add that any use case requiring minimal latency simply needs to stream audio up to the IBM Watson STT API over a Websocket -- the currently-supported approach involving recording the utterance then sending it simply doesn't cut it!