EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
DeepMind Blog·19d ago·5 min read
Gemini 3.8 text-to-speech says hello
TL;DR
DeepMind has launched Gemini 3.8, introducing two advanced text-to-speech models: Flash TTS for creative voice design and Flash-Lite TTS for efficient audio production. These models enhance audio experiences across various applications, enabling developers to create unique and expressive voices.
✦ Why It Matters
Explore Gemini 3.8 to enhance your audio projects with customizable and expressive voice generation.
Key Takeaways
- Explore voice customization with Gemini 3.8 Flash TTS for unique character voices.
- Utilize Flash-Lite TTS for efficient high-volume audio production.
- Leverage voice replication to create consistent vocal profiles from short samples.
- Access Gemini API for integrating TTS capabilities into your applications.
Full Summary
Gemini 3.8 expands the Gemini Audio family with two new text-to-speech (TTS) models: Flash TTS and Flash-Lite TTS. Flash TTS is designed for creative applications, allowing users to generate unique voices using natural language prompts, ideal for gaming and interactive media.
In contrast, Flash-Lite TTS focuses on high-volume, cost-effective audio production, suitable for dubbing and voice agents. Both models support over 100 languages and provide granular control over voice characteristics, pacing, and tone.
They also feature voice replication capabilities, enabling users to recreate voices from short audio samples while ensuring consent and transparency through built-in safeguards. The models have achieved top rankings in voice design and quality benchmarks, making them powerful tools for developers and enterprises.