Google's new Gemini TTS model is amazing! It can directly control the speaker's gender, tone, intonation, pronunciation of specific words, and age using prompts. It can control almost anything you can think of, without needing to switch to a separate speech model! My AI interactive comic book app is finally complete! It uses Nano Banana Pro to dynamically generate images for each scene, and the new TTS to generate unique voices for each character. twitter.com/op7418/status/1999...

When speaking Chinese, some old issues remain, such as a slight foreign accent. However, it's much better than previous TTS systems for speaking Chinese. Hopefully, they can fix this strange intonation problem in Mandarin.

I've adjusted the prompts; try it again if it sounded wrong before: