Loudness Standards: Why -14 LUFS?
Prevent platform compressors from squashing your sound with -14 LUFS integrated and -1 dBFS true peak QA.
Use motionmind music to create beat-synced background tracks with dynamic sections and auto-ducking.
A short promo needs music that starts and ends on time, changes where the story changes, and belongs to you. motionmind music writes one from code: a 48 kHz stereo WAV with no samples in it, built from the length and sections you ask for. The agent runs it during the build; you can run it yourself.
motionmind music --out src/assets/audio/music.wav --duration 12 --bpm 120 \
--sections "0:tense,2:groove,4:lift,5:outro" --bells "8,10" --duck 6:0.3
| Option | What it does |
|---|---|
--duration | length in seconds |
--bpm | tempo, 120 by default |
--sections "bar:kind,…" | where each section starts, counted in bars from 0 (not seconds) |
--bells "s,…" | a rising four-note motif at those seconds: a reveal, the logo, the CTA |
--duck s:depth,… | dips the bed around a hero sound or a voice-over line (depth 0.2–0.5) |
--seed | which variation you get; 817 by default |
The same command always writes the same file; in the clip, two runs give identical hashes. That's what keeps a version reproducible when it's rendered again. Want a different bed with the same structure? Change the seed:
motionmind music --out music.wav --duration 12 --seed 42 --sections "0:tense,2:groove,4:lift,5:outro"
This clip has sound: press play. It's the bed from the command above. At 120 BPM a beat is 0.5 s and a bar is 2 s, so 4:lift starts at 8 s. Plan scene changes on bar lines, or at least on beats, and they'll feel intentional.
Five kinds of section:
| Kind | Sounds like | Use it for |
|---|---|---|
tense | minor pad, ticking hi-hats | the problem |
calm | pad and soft keys | explaining, under a voice-over |
groove | major, kick, clap and hats | the product |
lift | a busier groove | the climax |
outro | an open chord, no drums | the CTA |
The bed goes into the version as an audio clip:
<audio id="music" class="clip" src="assets/audio/music.wav"
data-start="0" data-duration="12" data-volume="0.5" data-track-index="10"></audio>
The WAV is peak-normalised, not loudness-normalised. The final level comes from the master gain on the composition, measured on the render; see Loudness Standards: Why −14 LUFS?
Usually you just describe the music you want:
Continue demo/launch-6s: make v6 with a calmer bed (calm for the first 4 s, then groove)
and a bell on the CTA. Keep everything else.
Prefer your own track? Copy it into src/assets/audio/ and note its licence in the campaign's manifest.json.
Prevent platform compressors from squashing your sound with -14 LUFS integrated and -1 dBFS true peak QA.
Organize vector SVG logos, 2x captures, and 22 built-in CC0 sound effects in shared/brands/ for your agent.
Understand how MotionMind equips AI agents with a studio-grade video production process and versioned records.
Open your terminal, launch motionmind ui, and brief your AI agent.