Add Configurable silence between audio chunks

[bi1101]

No description provided.

[matatonic]

How does this work for you? Satisfied?

Given that the most common user right now is open-webui which also splits into sentences before sending to openai-speech, I'll probably add the silence pad to all output, not skipping the last.

[matatonic]

Can you add a command line option also, which takes the environment variable as the default? If you don't get to it, I will. Also - can you make the PR against the current dev branch instead? It's soon to be released.

[bi1101]

It works well.

I think addding padding to all sentences could be a better all-rounded approach for most usecases

[thiswillbeyourgithub]

How does this work for you? Satisfied?

Given that the most common user right now is open-webui which also splits into sentences before sending to openai-speech, I'll probably add the silence pad to all output, not skipping the last.

Hi, FYI since last release (0.3.16 i think) of openwebui we can set the tts split to be based on punctuation, paragraph or don't split