How to optimize AI voice quality when using Genny API in a multilingual environment

Hi everyone,

I am trying to integrate Genny API with a multilingual audio content creation project using That's Not My Neighbor Lovo AI. The problem I encountered is: when converting Japanese or Spanish text to AI voice, the sound quality and accent are sometimes unnatural, leading to a less smooth listening experience.

I have tried:

Using different Lovo AI voice presets

Adding punctuation, special characters to improve intonation

Adjusting speaking speed and pitch

But the results are still inconsistent, especially with languages with complex accents or syllables.

Does anyone have experience optimizing Genny API for multilingual voice? Are there any tips or best practices to improve naturalness?

Should I use SSML or special scripts for each language?

Anyone who has tried to create a multilingual batch voice while still keeping it natural, can share the setup?

I look forward to receiving shares from the community! I believe that if I can do it, it will help many people take advantage of Genny API and Lovo AI more effectively.