AudioLDM AI sound generator
AudioLDM generates ambience and effects up to 30 seconds, as MP3 or WAV. Use it for rooms, weather, and beds. Compare the credit prices below.
Best for
AudioLDM is the ambience generator on the Framings sound-effects screen. Use it for weather, rooms, crowds at a distance, and other beds that need to sit under a picture. Clips run up to 10 seconds or up to 30 seconds, as MP3 or WAV.
Controls and pricing
Describe the space more than the plot. Rain on a roof with no voices, or a museum with one set of footsteps, stays coherent. If you need a sharp one-shot under 22 seconds, ElevenLabs SFX is the other model on this screen and uses a different credit price.
How to use it
Open sound effects, choose AudioLDM, pick the length and the format, and generate. The table shows 10 and 30 seconds on the Free plan. A very short cue still has a floor, so it does not collapse to a token credit.
Good to know
AudioLDM is not a music model. A melody or a vocal belongs on Stable Audio or MiniMax Music. You can lay an AudioLDM bed under either of those in the video editor, and each generation keeps its own credit charge.
Where it fits
Thirty seconds is the length to use under a shot that holds. Ten seconds is the length for a transition. Export WAV if you will mix the bed with voice or music, and MP3 if the file is going straight into a quick cut. Name the space, the distance, and what should stay out of the recording.
Prompting tips
Leave voices out of an ambience prompt unless a distant murmur is the point. A clean room tone is easier to mix under dialogue you will add later. AudioLDM's 30 second WAV is the file to keep when the edit still has levels to set. Ten seconds is enough when the shot is a cutaway.
Credits and controls
- 10 seconds
- 30 credits
- 30 seconds
- 30 credits
Example prompts
Prompt idea
Distant rain on a tin roof, soft thunder far away, no voices, 30 seconds.
Prompt idea
Quiet museum room tone with a single pair of footsteps crossing stone.
Questions
How long can a clip be?
Up to 10 seconds or up to 30 seconds.
Which formats are available?
MP3 and WAV.
How is it different from ElevenLabs SFX?
AudioLDM goes to 30 seconds and offers WAV. ElevenLabs SFX stops at 22 seconds and exports MP3.
Can it write a song?
No. Use Stable Audio for instrumentals or MiniMax Music for vocal tracks.
Where are the credit prices?
The table lists 10 seconds and 30 seconds on the Free plan.