Background music for ads: brief the gaps, not the sound

An empty lyrics box means instrumental. That single decision is what separates a bed from a song.

Type the description in your own words. Comma-separated fragments work well: warm acoustic guitar, gentle, no drums, slow tempo around 70 BPM.

Suggestion chips write into the box rather than replacing it, so you can start from a genre and add the specifics.

Pick 15, 30, 60, 90 or 120 seconds. The price is flat at 20 credits at every length.

In the ad editor the music controls are turn it on, pick the track, and set how loud it sits.

There is no ducking, no stem separation and no tempo slider, so the description is where all the control lives.

A spoken voice occupies roughly 500 Hz to 4 kHz, and consonants live at the top of that range.

Anything in the track sitting in the same band eats those consonants, and the viewer stops following the words without knowing why.

So describe what should be absent. Sparse. No pads. Nothing in the midrange. Low end and high percussion only.

That reads like an odd brief and it produces a far more usable bed than anything asking for warmth, fullness or energy.

Ask for slow, around 70 BPM, when the voice-over is fast. Two fast rhythms fight and neither one wins.

The one word to avoid entirely is big. It reliably returns a wall of midrange.

A phone speaker reproduces almost nothing below roughly 500 Hz, and most ads are watched on one without headphones.

So the weight and warmth in your bed are inaudible at the point of contact. Only the midrange survives.

Which means the part that survives is exactly the part competing with the voice-over.

That is the reason a track that sounded fine on monitors turns to mush on a phone, and it is not a mixing skill problem.

Check every ad on a phone at half volume held at arm's length before it goes out.

If the words stop being clear, generate a sparser track for another 20 credits rather than pulling the level down until the music is pointless.

Music is a default rather than a decision for most people making ads, and it is often the wrong default.

A dense voice-over with no pauses gains atmosphere and loses clarity, which is a bad trade on a direct response ad.

A demonstration where the product makes a sound is stronger with that real sound than with a bed over it.

Play the cut muted and then with the voice only. If the voice-only version holds, music is optional.

If it feels flat in the gaps, that is the exact job for a bed, and you now know what to describe: fill the gaps, stay out of the words.

Twenty credits is small, but a track that makes an ad worse is expensive in a different currency.

Teams that run this test regularly end up using music on about half their ads rather than all of them, and their voice-overs get noticeably clearer.

That is a result nobody expects from a music module, and it is the most useful thing this page has to offer.

What this does is narrow and worth stating plainly. It turns a sentence about a sound into a usable bed in one step.

The price is small enough that a wrong answer costs nothing. That makes it a description problem rather than a production problem.

For somebody whose actual job is writing ads, that is a much better problem to have.

Keep a short list of the descriptions that worked. Three good sentences cover most of what a brand ever needs behind a voice.

Five plain limits sit around it. There is no automatic ducking under the voice. The level is one setting for the whole ad.

There are no stems, no tempo slider and no way to edit a returned track. You generate again for 20 credits.

No track runs past 120 seconds, so anything longer is assembled elsewhere.

A specific recognisable song is a rights negotiation rather than a generation problem.

And frame-accurate audio work belongs to a full audio editor. This is not competing for that job.

One last thing. The track is generated for you rather than licensed from a catalogue. Confirm the usage terms that apply to your own account, and keep the description and the date.