Spatial audio prompt builder
Flat AI-generated audio often sounds like everything is glued to the center of the speakers. Adding spatial descriptors — where a sound sits, how far away it is, and what room it lives in — produces audio that feels placed in real space. This builder assembles a clean prompt from a sound source plus position, distance, and environment cues.
How it works
Three layers create a sense of space. Stereo placement (hard left through center to hard right) sets the horizontal position. Distance is conveyed through loudness, brightness, and how much reverb wraps the source — distant sounds are quieter, duller, and wetter. Environment defines the acoustic space itself: a dry vocal booth has almost no reflections, while a concert hall or cathedral adds long reverb tails. The builder joins your source with these descriptors in a natural order the model can parse.
Tips and examples
- “Footsteps on gravel, far left, distant, in a large empty warehouse” places a faint, echoing source to one side — useful for tension and depth.
- “Whispered voice, center, very close, dry studio” sits intimate and forward, with no room around it.
- Match level to distance. Don’t describe a source as both “distant” and “loud and present” — pick one spatial story.
- One environment per prompt. Mixing “outdoor field” with “concert hall reverb” confuses the model; choose the space that frames the source.