Background music for commercials: generate a bed that fits

Background music for a commercial is an instrumental bed under the voiceover that ends on time. How to brief, generate, cut and duck one with AI.

5 min readSume
All posts

Background music for a commercial is an instrumental bed that sits under the voiceover, carries the brand's mood, and ends exactly when the spot does. You can license a stock track, or generate one from a brief written for the ad: instrumental, a set tempo and mood, the spot's length, and a lift at the call to action.

On Sume, the Music Router generates one track per request at $0.125 per audio generation, plus a 5.5% agent fee by default. The brief format comes from Music 1.0, and cutting and mixing from Timeline audio and Timeline 1.0, read on 2026-09-29.

What should a music brief for an ad say?

The Music docs suggest a brief with seven axes; AI music prompt examples has the general template. They are creative directions, not guaranteed output settings, so listen to every take. For a commercial, each axis has a job:

  • Say the length in the prompt: "a 30-second track". The request has no duration field; duration and duration_seconds are rejected.
  • Put exclusions in the prompt itself. A non-empty negative_prompt returns 400.
  • Close with "Instrumental, no vocals." and add "no spoken word" when the bed runs under narration, as the docs suggest.
Axes from Music 1.0, read 2026-09-29. The ad column is guidance, not a Sume setting.
AxisIn an ad brief
EmotionThe brand's mood, stated precisely: "proud and nostalgic", not "happy"
Genre / lineageOne style the audience recognizes
Tempo as a numberA BPM, so every cutdown has the same pulse
Key and modeMajor or minor sets the tone under the read
Instruments (2–4)Few, so the voice keeps the midrange
Arc with one named momentThe lift timed to the call to action, e.g. "full return at 0:24"
Era / productionMatches the look of the picture
curl -X POST https://api.sume.com/v1/music-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: spring-sale-bed-30s-001" \
  -d '{
    "prompt": "Bright, confident indie pop, 112 BPM, G major. Muted electric guitar, light claps, warm bass. Sparse under the first lines; full lift at 0:24 for the offer. A 30-second track. Instrumental, no vocals, no spoken word."
  }'

How do I make 15 and 30 second cutdowns?

Either brief each length separately ("a 15-second track"), or cut a longer take. A Timeline audio split slices a Sume-hosted file into up to 20 ranges and returns each as its own file, WAV (pcm_s16le) by default or MP3, at $0.01 per job. A split ends where you cut it, so let the render fade the bed out rather than stopping it dead. Generate music for a video with AI covers fitting a track to a finished cut.

How do I keep the music under the voiceover?

In a Timeline 1.0 render, the voiceover is the audio spine and the generated track goes in the optional soundtrack. In current code the render takes sound only from the spine and the soundtrack, so any sound on the video clips is dropped. How loud the bed should sit is covered in How to mix voice with background music. The soundtrack settings:

  • gain_db sets the bed's level.
  • duck_db from 0 to 20 lowers the bed under the spine; it needs a real spine, not silence.
  • fade_out_seconds up to 10 fades the bed at the end.
  • loop is also a documented soundtrack option.

How much does background music for an ad cost?

Each generation is one flat charge of $0.125 per audio, whatever the prompt length, and every take counts. A failed job releases or refunds the reservation. Take counts below are example assumptions.

Computed from the Music generation row on API pricing ($0.125 per audio) and the Timeline audio rate ($0.01 per job), read 2026-09-29. Before the 5.5% agent fee.
ExampleGenerationsSplitsPrice
One 30-second bed, first take kept10$0.125
One bed, 5 takes to choose from50$0.625
30 s bed, 5 takes, plus a 15 s cut51$0.635
A 15 s and a 30 s bed, 5 takes each100$1.25

Can I use AI-generated music in a commercial?

It depends on the terms of the tool you generate with. The Sume Terms of Service say Sume does not claim ownership of generated outputs created for your account, that paid plans include commercial use as described on the pricing page, and that you are responsible for reviewing outputs before commercial use. That is what the terms say, not legal advice. AI video generator for commercial use covers the same clauses for video.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume