HiAPI
Showcases
Lyria 3 Pro Parameter Panorama: A Three-and-a-Half-Minute Three-Part Score

Lyria 3 Pro Parameter Panorama: A Three-and-a-Half-Minute Three-Part Score

Taking Lyria 3 Pro's "parameter panorama" as the topic, this piece uses a real generation pipeline to demonstrate how to translate model capabilities into an executable prompt structure: use timestamps to mark the intro, verse, chorus, and bridge, specify genre, instrumentation, mood, and tempo, and overlay an atmosphere reference image to guide the creative direction—according to the schema, image_urls only affect mood and theme and do not become the visual source of the audio. The work itself is a complete track of about 2 minutes 30 seconds, 48kHz stereo, with clear section starts and ends, suitable as a selection sample for video soundtracks, game background music, and brand short films. For developers, the focus is the reproducible input structure: the prompt carries the section map, the seed provides an anchor for repeatable generation, and the reference image supplies semantic cues; the model family description states that this version outperforms Lyria 3 in prompt adherence and structural understanding, with a single generation limit of about 3 minutes, and overly long requirements need to be spliced in segments before post-production.

Style Structured arrangement · Text-to-Music · Soundtrack

Model Lyria 3 Pro

Reference inputs

  • image_urls

    image_urls

Prompt

Try this model
Instrumental cinematic and electronic hybrid style, 96 BPM, 48kHz stereo, total length about 2 minutes 30 seconds. Warm analog pads and felt piano form the intro, followed by soft brush percussion; muted arpeggios carry the verse while the low end stays restrained; the chorus opens with broad strings, a sustained sub bass, and a simple four-on-the-floor groove, bright but not aggressive; the bridge contracts into a piano solo and reversed-sound texture, then the final chorus resolves into a brief decaying outro. Mood: hopeful, steadily driving, spacious. Clean dynamic range, no distortion, no vocals.

Copy the full prompt and adapt it for your own generation.

Tags