Google has released Lyria 3.5, a music generation model that allows users to edit specific sections of a track without restarting the creation process. The update is available through Google Flow Music and introduces a feature called Selective Section Painting. This tool enables users to modify individual parts of a song or expand short melodies into full compositions. The model also permits fine-tuning of tempo and duration for vocals, drums, bass, and other elements. Generated tracks can range from thirty seconds to three minutes in length. Google states that Lyria 3.5 produces more natural-sounding melodies, improved lyrics, and realistic vocals with clearer pronunciation. When the previous version launched, the company noted it trained the model on materials used under YouTube and Google terms, partner agreements, and applicable law. Google did not provide specific details about the training data for Lyria 3.5 and has not yet responded to further inquiries regarding this information.
The ability to edit sections without restarting addresses a common frustration with current generative audio tools, which often require users to discard entire outputs for minor adjustments. This workflow improvement reduces time wasted on repetitive prompting and allows for more iterative development of musical ideas. The precise control over individual instrument tracks offers a degree of flexibility previously unavailable in automated music generation.
* Tracks can run for between 30 seconds and 3 minutes.
* Users may adjust tempo and duration for specific instruments.
* Selective Section Painting allows editing without restarting generation.
