YEONJU KANG Azure Speech TTS uses neural voices that are trained to recognize punctuation marks and other special characters. The usage of these characters will ensure the voice sound as natural as possible i.e if you use an comma the voice that is being used will pause briefly and similarly full stop and other notable characters. This is the reason why the characters are also counted for billing purposes and using them will ensure your text is synthesized as naturally as possible. The service also provides an option to customize the behavior as much as possible if you are using SSML as input for TTS. Primarily, you can adjust the pitch, volume, rate, speaking style, prosody and other various parameters to ensure the usage of characters compliments them. An easy way to test this out is by using the speech studio, try the studio audio content creation tool with the defaults and by adjusting the rates using the slider. This will help you understand how the parameters compliment usage of these characters. Please see this page for details about their usage. I hope this helps!!
If this answers your query, do click Accept Answer
and Yes
for was this answer helpful. And, if you have any further query do let us know.