Google Lyria 3.5 enables surgical music editing
google released lyria 3.5, a music model offering 'selective section painting' to edit track parts without regeneration, streamlining ai music production workflows.
google deployed lyria 3.5, an upgrade to its music generation model, featuring improved vocal pronunciation, natural melodies, and support for tracks up to 3 minutes. available through google flow music, the model adds selective section painting, giving creators granular control to modify specific stems or sections without reconstructing the whole track. the tool also allows precise adjustment of vocals, drums, and bass, reducing iteration time for producers. despite these advancements, google did not disclose the training data sources for the new model. this release signals a shift from pure generation to controllable editing in generative audio.