We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
DeepMind Blog·Feb 18·5 min read
A new way to express yourself: Gemini can now create music
TL;DR
Many people find it challenging to create music without formal training or tools. Gemini has introduced Lyria 3, a music generation model that allows users to create 30-second tracks by describing ideas or uploading images.
✦ Why It Matters
Engineers can leverage Lyria 3 to develop user-friendly music generation tools that democratize creativity.
Key Takeaways
- Use text or images to generate custom music tracks.
- Explore creative control over style, vocals, and tempo.
- Share your unique tracks easily with friends.
- Verify AI-generated content using SynthID watermark.
How It Works
Users input a description or upload a photo, and Lyria 3 generates a 30-second music track based on that input. The model automatically creates lyrics and allows users to specify musical elements like genre and tempo, resulting in a personalized audio experience.