Lyria 3

Google's music and ambient-audio model family. Two variants share the same prompt surface but trade off length for runtime:

• Lyria 3 Clip — short clips, loops, and previews up to 30 seconds (~30s runtime).
• Lyria 3 Pro — full-length songs with verses, choruses, and bridges up to ~2 minutes (~1min runtime).

Both produce 44.1 kHz stereo audio.

Capabilities

Feature	Lyria 3 Clip	Lyria 3 Pro
Text-to-Audio	Yes	Yes
Max Duration	30 seconds	~2 minutes
Output	44.1 kHz stereo (MP3 / WAV)	44.1 kHz stereo (MP3 / WAV)
Vocals + instruments	Yes	Yes
Reference audio	No	No
Negative prompt	No	No
Reference image	Yes (up to 10)	Yes (up to 10)

When to use which

• Lyria 3 Clip — fast iteration on a single musical idea, SFX-adjacent loops, short cues for stings or transitions.
• Lyria 3 Pro — production-ready music with full song structure (intro → verse → chorus → bridge → outro), or longer ambient pads / soundscapes.

Prompt phrasing carries over between the two — start with Clip to lock the genre, instrumentation, and mood, then re-run on Pro for the full piece.

Prompting guide

Good prompts are descriptive and specific. Start with the core musical idea, then refine with keywords across the dimensions below. Lyria 3 reads everything in your prompt as a single brief — there are no separate fields for tempo, key, instruments, etc., so encode them inline.

Core dimensions

Element	What it controls	Examples
Genre & Style	Primary category + stylistic subtype	electronic dance, cinematic orchestral, lo-fi hip-hop, 80s synth-pop, Berlin techno
Mood & Emotion	Affective register	energetic, melancholy, peaceful, tense, nostalgic, dreamy
Instrumentation	Which instruments lead vs. accompany	solo piano, Fender Rhodes, string orchestra, electronic drums, acoustic guitar
Tempo & Rhythm	Pace + rhythmic character	120 BPM, slow ballad, driving beat, syncopated rhythm, gentle waltz
Soundscape (optional)	Background texture / environment	rain falling, vinyl crackle, spacious reverb, underwater feel
Production quality (optional)	Mix / recording style	high-quality production, clean mix, vintage recording, raw demo feel

Song structure (Lyria 3 Pro)

Pro understands musical structure and will follow your outline. Use any of:

Section tags — [Intro], [Verse 1], [Pre-chorus], [Chorus], [Bridge], [Outro].

Arrow flow — [Intro] → [Verse 1] → [Chorus] → [Verse 2] → [Chorus] → [Bridge] → [Outro].

Energy / intensity arc — Gradual crescendo throughout, adding one instrument at a time, Build tension in the pre-chorus, drop to silence before a massive chorus.

Timestamps — exact-second control:

[0:00 - 0:12] Intro: Begin atmospherically with just Fender Rhodes chords,
              warm reverb, light atmospheric texture. Intensity: 1/10
[0:12 - 0:24] Verse 1: Drums enter (kick + snare). Synth pad swells in
              the background. Intensity: 3/10
[0:24 - 0:36] Build: Add syncopated hi-hat and a warm rounded-synth lead
              with a hopeful tension. Intensity: 5/10
[0:36 - 0:48] Chorus: Pull intensity back; return to core Rhodes groove.
              Intensity: 4/10
[0:48 - 1:00] Outro: Drums and bass drop out; Rhodes + lingering pad
              fade. Intensity: 2/10

You can also drop in instant cues anywhere in the prompt: Build to a drop at 12s, The chorus kicks in at 22s, Someone says "what" every 2 seconds.

Lyrics

By default Lyria generates vocals + lyrics matching the genre and mood. Three ways to steer:

Your own lyrics — include them with section tags:

Lyrics:

[Verse 1]
Walking through the neon glow,
city lights reflect below.

[Chorus]
We are the echoes in the night,
burning brighter than the light.

Parenthesize for echoes / backing vocals: "Let's go (go)".

Steer Lyria's own lyrics — describe the subject: "The lyrics are about lost love and the pain of heartbreak. A powerful chorus focuses on getting over the pain and moving on."

Instrumental only — say it explicitly: "Instrumental only, no vocals". Without this, Lyria may add vocals even when the genre doesn't normally have them.

Language — Lyria generates lyrics in the language you write your prompt in. Write in French, get French lyrics. Or instruct explicitly: "Write the lyrics in French".

Vocal effects — non-lyric vocal moments are fair game: "A repeating sample says 'I can't believe this!' throughout the song", "The track opens with a conversation about movies before segueing into the pop song".

Vocal profile

For sung vocals, give a detailed singer profile — gender, timbre, range:

• Female Soprano — clear, crystalline timbre with agile soaring quality; airy breathy whistly highs
• Female Alto — rich, warm, husky lower range; smoky timbre with vocal fry, soulful and resonant
• Male Tenor — bright, piercing, energetic; youthful timbre with a slight nasal edge; high belting power
• Male Baritone — deep, chocolatey, velvet-smooth; resonant chest voice with crooning delivery
• Weathered Rocker (male) — raspy, gravelly, reminiscent of 90s grunge; strained upper range

Reference images (multimodal)

Both models accept up to 10 reference images alongside your text prompt. Use @-mentions in the inspector to reference uploaded artifacts inline:

ambient track with sunset feel from @[Image 1] transitioning to
nighttime rave energy like @[Image 2]

Behind the scenes the inspector compiles each mention into a positional reference so Lyria associates each image with the prompt fragment that introduced it. Images NOT mentioned are still passed in (as overall mood context) but without positional anchoring.

Key, BPM, intensity

These are first-class controls — Lyria honors them when stated:

• Key / scale — in G major, D minor, chromatic
• BPM — 120 BPM, tempo around 85 BPM
• Intensity rating — Intensity: 5/10 (Medium) inside a structured prompt; or relative phrasing: quiet piano intro, build to a massive wall of sound, then drop to silence

Worked examples

Lofi study beat:

A 30-second lofi hip hop beat with dusty vinyl crackle, mellow Rhodes piano chords, a slow boom-bap drum pattern at 85 BPM, and a jazzy upright bass line. Instrumental only.

Upbeat pop with vocal hook:

An upbeat, feel-good pop song in G major at 120 BPM with bright acoustic guitar strumming, claps, and warm vocal harmonies about a summer road trip.

Cinematic trap underscore:

A dark, atmospheric trap beat at 140 BPM with heavy 808 bass, eerie synth pads, sharp hi-hats, and a haunting vocal sample. In D minor.

80s synth-pop retro hit:

A 1980s-style synth-pop track with a driving beat, shimmering synthesizers, and a catchy anthemic chorus. Retro-futuristic, reminiscent of classic 80s pop with modern production polish. Upbeat and danceable, around 120 BPM. Verse-chorus structure with a memorable instrumental hook. The lyrics are about the feeling of getting ready for a party.

Limitations

• No reference-audio input — drive the model with a prompt and an optional reference image
• No seed / determinism — each generation is fresh, even with the same prompt
• No negative prompt — describe what you want, not what to avoid; use "Instrumental only, no vocals"-style positive framing instead
• No multi-turn editing — generation is single-shot; can't refine in place
• Safety filters block specific-artist-voice impersonation and copyrighted-lyric requests
• All output carries a SynthID audio watermark (imperceptible to listeners)