Add visual feedback for spoken content

[Clstroud]

While spoken content is being synthesized, two improvements can be made:

1. The keyboard key that speaks typed content should animate its SFSymbol to indicate that content is being spoken
2. The textual content being spoken should highlight the portion of text that corresponds to the current utterance while the synthesizer is speaking. This creates a karaoke-style effect and visually reinforces what is happening.
3. This would also apply to collection view cells outside the keyboard.