Google introduced the music model Lyria 3 in Gemini: users can create 30-second tracks
Google announced the launch of the new music model Lyria 3, integrated into Gemini. From now on, users can create 30-second tracks based on a text query or remix existing compositions. The tool will also be available in the Dream Track feature on YouTube, where it can be used to generate musical accompaniment for Shorts.
What's known
Lyria 3 allows you to create music not only based on a textual description but also on an image or video. If necessary, the user can further adjust the parameters, including changing the tempo, drum style, or overall mood of the composition. Covers created using the Nano Banana image model can be added to tracks. The system can also automatically generate song lyrics.
Google claims that Lyria 3 creates more realistic and musically complex tracks compared to the company's previous audio models. Currently, the results are limited to 30 seconds, but in the future, the duration may increase, and the tool may appear in other company services.
All generated tracks are marked with the SynthID digital watermark, which allows the identification of content created by artificial intelligence. Access to the feature is open to users aged 18 and over who use Gemini in English, Spanish, German, French, Hindi, Japanese, Korean, or Portuguese.
Source: Google
