What is Lyria 2?

Lyria 2 is a high-fidelity instrumental music model exposed on Oakgen.ai. It's tuned for the kind of music that lives under video and editorial work — orchestral cues, ambient pads, lo-fi beds, electronic soundscapes — rather than chorus-driven songs with lyrics. The interface is intentionally tight: a prompt describing the track, an optional negative prompt for what to keep out (vocals, distortion, low quality), and an optional seed for deterministic results. Output is delivered as a clean WAV file at high quality, which makes it the right pick when the audio is going into a video edit, a game soundtrack, or a long-form piece that needs to hold up at full bitrate.

Why Lyria 2 is popular

• Audio fidelity is the headline — Lyria 2 returns clean, full-band WAV tracks that hold up on real monitors instead of just laptop speakers.
• Negative prompts give you precise control over what to exclude (no vocals, no distortion, no slow tempo), which is a missing knob on most instrumental music models.
• Seed control means you can lock a generation and explore variations from a known starting point — critical when you finally land the take you want and need to iterate without losing it.
• Strong on cinematic, ambient, classical, and lo-fi territories where tonal coherence matters more than vocal hooks.
• Comfortable default on Oakgen.ai when a video edit needs a clean instrumental bed without a song's chorus pulling focus.

When to use Lyria 2

• You're scoring a video edit and need an instrumental bed, not a track with vocals.
• You want a cinematic or ambient cue with clean orchestration and high tonal coherence.
• You need to lock a specific sound and iterate around it — set a seed, then run small prompt variations from that anchor.
• You're producing game audio, meditation tracks, or long-form editorial cuts that demand clean WAV output.
• You want explicit control over what to keep out of the track (vocals, distortion, tempo issues) via a negative prompt.

How to use Lyria 2

1. 1

   Describe the track or write lyrics

   Enter a prompt describing genre, mood, tempo, and instrumentation — or supply your own lyrics for a vocal track. You can combine both for precise control over the output.

2. 2

   Set duration and style

   Choose clip length (typically 30 seconds to 4 minutes depending on model), vocal style, and any model-specific controls like reference tracks or instrumental-only mode.

3. 3

   Generate

   Hit generate. Music models usually return in 1–3 minutes with a complete, structured track — intro, verse, chorus, and outro — in a single file.

4. 4

   Download or extend

   Export MP3 or WAV for use in videos, podcasts, games, and creator content. Many models support extension — take a promising 30-second clip and extend it into a full track.

Tips for better results

• Be specific about genre and mood. 'Lo-fi hip-hop with a jazzy piano and lazy beat' gets you much closer than 'chill music'.
• Name instruments explicitly. 'Acoustic guitar, soft strings, subtle synth pad' produces a far more coherent track than abstract descriptions.
• Describe the arc: 'builds from sparse intro to full chorus at 0:30, then drops at 0:45'. Structure cues help models produce tracks with musical narrative.
• Tempo matters. Include BPM if you have a target ('90 BPM downtempo') or descriptive cues ('slow, ballad tempo' vs 'energetic, driving 120 BPM').
• For vocal tracks, give the model a clear lyrical theme and emotional tone — 'melancholic breakup ballad' vs 'confident anthem about persistence'.
• Reference real artists or eras carefully — some models respond well to 'in the style of 70s Motown' but avoid copying specific living artists for commercial use.
• Iterate by re-running with small prompt changes. Same seed + tweaked genre cue often produces related variants you can stitch together.

Popular use cases

Lyria 2 vs other motion-transfer tools

vs Vocal-led song generator

Vocal-led song models on Oakgen.ai produce full songs with a chorus, harmonies, and lyrics. Lyria 2 produces an instrumental bed. Pick Lyria 2 when the track needs to sit under a voiceover or video edit and not pull focus; pick the vocal model when the song itself is the content.

vs Short-clip instrumental generator

Fast-clip instrumental tools on Oakgen turn around 30 seconds of audio in a couple of seconds and trade some fidelity for speed. Lyria 2 leans the other way — higher audio fidelity, cleaner WAV output, and explicit negative-prompt and seed control. Pick the fast tool when speed matters more than mastering; pick Lyria 2 when the audio has to hold up in a professional edit.

vs Stock instrumental music library

Stock libraries are quick but recycled — the same handful of cues turn up across countless brands and trigger Content ID matches when they're popular. Lyria 2 on Oakgen produces an original instrumental on your brief with full commercial rights on paid plans, so the same track can ship across every campaign without licensing follow-ups.

Frequently asked questions

Related tools on Oakgen.ai

Tutorials and guides