Design: Microphone input indications

User story

Today, the end-user is not "crispy clear" on whether we are picking up their voice on the microphone.
States:

• Dictate state (Started)
• Idle
• Starting
• Stopping
• (state between idle and starting) "Will start once synthesizing is complete" <-- this is not the final state name; any suggestions are welcome.

Alternatives

It is not trivial to customize the microphone button because:

• Deep customization is required
• Cognitive Services SDK does not fully support all events to build the UI
  • Developer will need to write their own package to listen to microphone

Other implementations

For example, Siri show a "bouncing sphere" to indicate when the user speak. And the bounce effect is based on microphone input, so the user is extremely clear that Siri is recording.

Potential implementation

Tomorrow, we could make the microphone record more prominent by following other apps.

In the future, we could add wake word capability. The end-user need a very clear indications that we are recording and analyzing their voice.

[Enhancement]

The issue names no files, tests, or entry points. Start by examining the microphone controls and speech-state handling, then clarify the proposed states and visual behavior; done would require an agreed design and implementation that clearly indicates recording, starting, stopping, idle, and deferred start.