@PAVAGEAU Perrine If you are using the voices from speech studio Audio content creation screen, the voice usage is billed only when you fire a request i.e TTS request for which the voice is generated on the number of characters that are synthesized. You need not close or suspend any resource if you do not use them. The usage is based on the number of characters whenever a request is posted to the service or endpoint. See the TTS FAQ for more details. 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.