A month after the debut of Lyria 3, Google said on Wednesday that it will be releasing a music generation model called Lyria 3 Pro. Unlike the 30-second songs available with the Lyria 3 model, the new model will allow users to produce tracks up to 3 minutes long.
According to the business, the Lyria 3 Pro model will provide improved creative freedom and customization in addition to enabling users to produce lengthier tracks. Because the model has a better understanding of track structure than its predecessor, users can additionally describe other sections of a musical piece in the prompt, such as intros, verses, choruses, and bridges.
Google previously added the option to create music to the Gemini app with the introduction of Lyria 3. The Gemini app is also launching the Pro model, but it will only be available to subscribers who have paid for it. Additionally, the company is releasing Lyria3 Pro for its Google Vids video-editing software and ProducerAI, a music-production tool powered by gen AI that Google purchased last month.
Additionally, Google is enhancing its enterprise tools with Vertex AI (available in public preview), the Gemini API, and AI Studio via the Lyria 3 Pro model.
Google stressed that it trained its model using data from its partners as well as acceptable data from YouTube and Google. Additionally, it was stated that the model does not imitate an artist. On the other hand, it stated that if users designate an artist in prompts, it uses that artist’s “broad inspiration” to create a song.
Every music made with Lyria 3 and Lyria 3 Pro has a SynthID to indicate that artificial intelligence was used in its creation.
In order to prevent AI crap makers from misattributing music, Spotify introduced new capabilities earlier this week that allow artists to review songs released under their name. In the meantime, Deezer has released technologies that enable any streaming service to recognize music produced by artificial intelligence.