Skip to content
Music model · Google

Lyria 3 when something has to go under the clip today

Most ads do not need a song, they need a bed that stays out of the voiceover’s way. This engine does exactly that, and it does it faster than the other four: describe the mood, the instruments and the tempo, and about fifteen seconds later you have a track with no stock licence attached to it.

Rendered by this model

Turn the sound on and listen

–:–– / –:––
A warm acoustic bed for a bakery advert: guitar, upright bass, a little cello. No vocals.This was generated by the model this page is about, exactly the way the app runs it. It is not a photo shoot and not client work.

Specification

Built by
Google
Starts from
a written brief · your own photo
Track length
the engine decides
Your own lyrics
does not take your own words
Vocals
steered by the brief, with no switch
Render time
about 15s
Cost
1 credit per track

What it is good for

Fifteen seconds, not three minutes
Measured across four probe runs: five to twenty seconds. You can audition three different moods in the time one singing engine takes to start.
Scores your own graphic
The only engine here apart from its Pro sibling that takes a picture. Attach a finished ad graphic and the track is written for that shot, rather than for a brief you still have to write.
Sits under a voice
Ask in the brief for space in the middle of the spectrum and you get it. A bed that fights the voiceover is worse than silence.

When to pick something else

It will not sing your words — it takes no lyric sheet at all. If the ad needs a verse and a chorus, that is MiniMax Music 2.6. It has no instrumental switch either: it publishes a negative prompt field and then refuses any non-empty value, so the ask has to go in the brief itself. It takes no length, so if the track has to be exactly thirty seconds you want ElevenLabs Music. And its content checker can be strict: it refused one of our Pro briefs outright without saying which word it objected to.

See every model

How to brief this model

The things this engine handles differently from the rest. The habits that hold for every model are collected separately, in the help centre.

  1. Describe the mood and the instruments, not a track title. "Calm piano and strings, slow tempo" works; "something like a perfume advert" does not.
  2. Write into the brief that it should have no vocal if the bed goes under a voiceover. This engine has no separate switch — a word in the brief is the only way.
  3. Upload the graphic if you want the music to match the frame. This is one of two music engines that take a photograph as a hint about mood.
  4. Use the render time. Fifteen seconds a track means you can audition four moods in a minute — the fastest music engine on the list.
  5. Do not expect sung lyrics or an exact length. MiniMax Music 2.6 is for words, and ElevenLabs Music for a track that has to be exactly 30 or 60 seconds.
How to write a brief that works

Questions about this model

Try this model on your own product

Sign up without a card and get 40 credits. Switching model is one click, so you can send the same brief through several of them and keep the best frame.