Google has released its Lyria 3.5 music generation model, which introduces a new feature called 'Selective Section Painting.' This allows users to edit specific parts of a track or transform short melodies into full songs without having to restart the process. According to Google, the model produces more natural-sounding melodies, better lyrics, and more realistic vocals with clearer pronunciation. Users can also control tempo and track length more precisely, with tracks running from 30 seconds to 3 minutes. The model is now available through Google Flow Music.
The 'Selective Section Painting' feature enables users to fine-tune the tempo and duration of vocals, drums, bass, and other elements. This capability enhances the editing process by allowing for targeted modifications rather than overhauling the entire track. Google did not provide details about the training data used for Lyria 3.5, but when it launched Lyria 3, the company stated it had trained the model on materials YouTube and Google had the right to use under their terms of service, partner agreements, and applicable law. Google did not immediately respond to questions about the training data for Lyria 3.5.
Google launched Lyria 3 in a manner that emphasized its training data sources, but the details for Lyria 3.5 remain unclear. The company has not provided specifics on the data used to train the latest version of the model. Source: thedecoder
