Crowd sound effects
Crowd sounds depend on the room. Say how big the crowd is and how far away the listener is.
Listen: 10 samples from the crowd collection
Every clip shows the exact prompt that produced it. The render script reads its prompts from the same file as this page. Copy one, change a word and generate your own.
Goal roar
via audiocpp-stable-audio-3-sfxfootball stadium erupting as a goal is scored, enormous crowd roar building and sustaining, huge open space
Boos and jeers
via audiocpp-stable-audio-3-sfxlarge crowd booing and jeering, sustained disapproval with individual shouts cutting through, arena
Airport terminal
via audiocpp-stable-audio-3-sfxbusy airport terminal ambience, distant announcements, rolling luggage and layered conversation, large hall
Playground
via audiocpp-stable-audio-3-sfxschool playground at break, children shouting and laughing with a ball bouncing, open outdoor space
Protest chant
via audiocpp-stable-audio-3-sfxprotest crowd chanting rhythmically with a megaphone cutting over the top, city street, echoing
Street market
via audiocpp-stable-audio-3-sfxoutdoor street market, vendors calling out over dense shopper chatter and cart wheels on cobbles
Theatre settling
via audiocpp-stable-audio-3-sfxtheatre audience settling before a show, murmuring, coughs and seat creaks, then hushing
Panic scatter
via audiocpp-stable-audio-3-sfxcrowd panicking and scattering, screams with running footsteps dispersing in every direction, street
Crowd gasp
via audiocpp-stable-audio-3-sfxlarge audience gasping in unison then breaking into surprised murmur, medium hall
Standing ovation
via audiocpp-stable-audio-3-sfxconcert hall standing ovation, thunderous sustained applause with whistles and cheers, large hall
Prompts that work
These models respond to physical description. Name the material, the action and the space.
Small murmur
small crowd murmur, 10 to 15 people talking indoors at low volume, indistinct overlapping speech, no individual words audible, room tone
A headcount ("10 to 15 people") keeps the murmur small. Without a number the model defaults to an ambiguous mid-sized crowd.
Stadium cheer
large stadium crowd cheering, hundreds of voices layered, sudden rise in energy, outdoor open space, 3 seconds
"Sudden rise in energy" is the moment a crowd reacts to a play. It separates a cheer from a steady ambient roar.
Restaurant bed
restaurant background chatter, dozens of overlapping conversations at moderate volume, occasional cutlery clink, indoor room tone, steady and even
A small detail like "occasional cutlery clink" places the scene better than "restaurant" alone, which tends to come back as generic chatter.
Angry mob
angry crowd shouting, dozens of voices layered with aggressive tone, outdoor open space, rising intensity, no distinct words
"Aggressive tone" changes the vocal texture, separating it from a cheering or murmuring crowd of the same size.
Applause
sudden crowd applause burst, indoor auditorium, dozens of hands clapping at once, sharp onset, 2 seconds, natural decay
"Sharp onset" plus "natural decay" give the shape of a real applause burst: instant start, gradual settle.
Making crowd sounds that fit
Size the crowd with a number
"Big crowd" and "small crowd" are relative, and the model has no fixed scale for them. Give a headcount or venue size ("10 to 15 people", "hundreds of voices", "stadium"). That sets the density of overlapping voices more reliably than an adjective.
Say what the crowd is doing
A crowd's activity changes its texture as much as its size does. Talking, cheering, shouting, applauding and murmuring are different acoustic events. Name the activity. If it is a reaction (a cheer, a burst of applause), describe the shape of the reaction ("sudden rise", "sharp onset").
Keep individual voices out of it
A crowd becomes a group of speaking characters once words are intelligible. For background crowd tone, say "no individual words audible" or "indistinct". Otherwise the model may render a few voices clearly enough to sound like dialogue, which pulls focus in a mix.
Why generate locally
The models run on your machine, which changes the cost, the privacy and what you can automate.
Offline after download
Weights are fetched once. After that, generation, the library and the API all work with the network off. Everything lives in one SQLite file and one folder of WAVs under ~/.soundbox, and nothing is uploaded.
Your agent can use it
Soundbox runs an MCP server in the same process, with tools for generating and listing sounds. Generated files come back as absolute paths, so an agent can copy them into a project.
Licences stated up front
Every model's terms are shown before you download it, and non-commercial weights need an explicit confirmation. Soundbox is MIT. The licence of the model that made a sound governs what you may do with it.
claude mcp add --transport http soundbox http://127.0.0.1:17494/mcpcurl -X POST http://127.0.0.1:17494/api/generate \
-H 'Content-Type: application/json' \
-d '{"model":"audiocpp-stable-audio-3-sfx","prompt":"8-bit coin pickup blip","duration_s":1}'Frequently asked questions
- Can I use these sounds in a commercial video or game?
- It depends on the model that made them. Soundbox shows each model's licence before you download its weights. The Stable Audio 3 models use the Stability AI Community License: commercial use is free below USD $1M annual revenue, and above that Stability requires an enterprise licence. ACE-Step 1.5 is MIT. MiniMax's models allow commercial use under a revenue cap and require an attribution notice. Some voice and stem-separation models are non-commercial only. Soundbox itself is MIT, but the audio is governed by the licence of the model that made it. Read that licence before you ship; this summary is not legal advice.
- Does this run offline?
- Yes, once the weights are downloaded. Soundbox runs the models locally and stores everything in one SQLite file and one folder of WAVs under ~/.soundbox. There is no account, API key or upload step. Only the initial model download needs the network.
- What hardware do I need?
- An NVIDIA GPU with 8 GB or more. CPU-only is supported, but each generation then takes minutes. That is fine for a few short clips and slow for a large batch.
- Can my coding agent generate these for me?
- Yes. While Soundbox is running it serves MCP at http://127.0.0.1:17494/mcp, with tools for generating and listing sounds. An agent can generate a sound as part of the task it is already doing and gets back an absolute file path to use in your project.
- How is this different from downloading a sample pack?
- A crowd pack gives you a fixed size and mood recorded at someone else's venue, and joining two different crowd clips usually exposes the seam. Generating lets you set the headcount, the indoor or outdoor space and the mood your scene needs, so a cheer and the murmur before it share the same voice character. It costs time and a GPU.