does the TTS pricing refer to the net voice generation time? or is it an estimate for the TTS costs of each minute/hour of conversation (regardless of which side is speaking)? https://cdn.discordapp.com/attachments/1211484481848873020/1286336105356197898/Screenshot_2024-09-19_at_16.39.40.png?ex=66ed895a&is=66ec37da&hm=6b7b2ccd1c9a34569a36308ca135e10873b4a99e130de86095f340d7364b40cd&