Google has unveiled two new text-to-speech (TTS) models, Gemini 3.8 Flash TTS and Flash-Lite TTS, marking a significant leap in AI voice generation capabilities. These models support over 100 languages and introduce innovative features that allow users to design custom AI voices from scratch using simple text descriptions.
Empowering Creative Voice Design
The most notable feature of these new models is the ability to craft entirely new voices based on textual prompts. For example, a user can describe a voice as 'warm and youthful' or 'deep and authoritative,' and the system will generate a corresponding audio output. This level of customization opens doors for content creators, filmmakers, and developers who need unique vocal tones for their projects.
Enhanced Flexibility and Control
Beyond voice creation, the models offer advanced control over dialogue. Users can add stage directions to individual lines, enabling nuanced performance adjustments. Additionally, the system supports two-voice dialogues generated from a single script, simplifying the production of conversations. A voice cloning function also allows users to build a voice profile from just a 30-second audio sample, making it easy to replicate specific speakers or characters.
Implications for the Industry
These updates position Google at the forefront of AI voice technology, offering more flexibility and creative control than many existing solutions. With applications ranging from audiobook narration to video game characters, the new Flash TTS models are poised to reshape how developers and content creators approach voice synthesis. As AI continues to evolve, tools like these may become standard in multimedia production, democratizing access to high-quality voice generation.