Strip silence from SAPI 5 utterances

[nicopn]

Is your feature request related to a problem? Please describe.

There are many SAPI voices that have silence at the beginning and/or end of their utterances, making them less responsive.

Describe the solution you'd like

When using SAPI, synthesize the text, but before sending it to the audio device, remove all silent or near-silent samples from the beginning and end. This will improve responsivity in those voices

Describe alternatives you've considered

None

Additional context

N/A

[seanbudd]

Closing as duplicate of #13284