1. No. The speak each character option is not using recorded system prompts, the tts engine is generating the audio on the fly.
2. You could split the string yourself and separate the characters with blank audio. You could also try a little manipulation; i.e. if it's an "a" play "aye" or "ay" instead.
3a. You could try using a different TTS engine or voice so see if you like it better.
3b. You could create user prompts for each character. Upload your own, good recordings. In the flow split the string, loop through each character, and play the corresponding user prompt for each one.
------------------------------
Melissa Bailey
Genesys - Employees
------------------------------