Integrate Watson API for speech-to-text

[jeffpaul]

Splitting this out from #2

[Sidsector9]

Additional info:

The IBM Watson Speech to Text provides 3 interfaces for speech recognition:

• The synchronous HTTP interface
• The asynchronous HTTP interface
• The WebSocket interface (recommended)

See more on Advantages of the WebSocket interface.

@jeffpaul This service provides a lot of features, few of them such as: Speaker labels, Profanity filtering and Background audio suppression, etc.

Can you expand on the use case of this feature? That way we can list out the features that can go with the implementation.

[jeffpaul]

@Sidsector9 I believe the original thought on this enhancements was taking live speech and generating text from that for captions, so the speaker labels bit you highlighted probably most applies here.

[jeffpaul]

Could similarly look at using OpenAI's Whisper for this.