Contextual capabilities
Lyria 3 introduces significant advances in creative autonomy. Unlike previous models, the system automatically generates lyrics based on the context provided by the user, also offering control over the style, voice and tempo of the track.
The experience is enhanced by multimodality, allowing Gemini to “read” photographs or videos to compose soundtracks that suit the visual environment.
In the YouTube ecosystem, this technology will be integrated into Dream Track, helping content creators raise the sound quality of their videos. Shorts. Each creation is accompanied by a unique cover art, generated by the Nano Banana image model.
Transparency and security ensured by SynthID
The North American company reinforced its commitment to responsible artificial intelligence by incorporate SynthID technology into all generated tracks. This digital watermark, although imperceptible to the human ear, allows the origin of the content to be identified, making it detectable as having been generated by AI.
For the first time, users can also use Gemini itself to verify the authenticity of audio files by uploading them to the app to detect the presence of this watermark.
Regarding copyright, the company guarantees that model training was carried out under strict partnership contracts and that the system is programmed for “original expression”, avoiding direct imitation of existing artists.
Leave a Reply