Lyria 3.5 Multi-Style Sectional Song Excerpt Demo

A demo of Lyria 3.5's text-to-music capability for developers: using a 90-second music description, three styles—city pop, ambient electronic, and light rock—are arranged into the intro, verse, chorus, and bridge of a single song, to verify how the model reasons about and plans song structure before generating audio, and how it transitions style and mood between sections. On the input side, the music description (prompt) field declared in the author's pipeline is used to describe genre, instruments, mood, and section structure, with the song title, BPM, and duration additionally provided to constrain rhythm and scale; the model outputs a 44.1 kHz stereo MP3 according to the official capability. The work contains no reference materials that require pipeline generation—all content is driven directly by this description—so the submit snapshot shown on the details page is the complete generation condition. Applicable scenarios: rapidly producing structurally complete, mood-adjustable soundtrack prototypes for short videos, podcasts, and interactive media; verifying whether "multi-style switching within a single song" holds before formal arrangement; and checking whether the model's section boundaries match expectations against the section labels and duration constraints in the description.