Feature: Real-time pronunciation feedback via transcription

[wcjord]

We currently give feedback via transcription after the message is sent but we should explore doing this in real-time.

[linhtphung]

This may be done in a number of ways:

• Live closed caption
• Sound waves indicate intonation
• Comparison with a sound waves from "native speakers"
• Allowing users to record their message again
  Whatever it is, this can be saved for the next phase of development