A costume is an argument with the room's visual attention; a sound cue settles it. The room hears the whoosh with the arm swing and the belief lands. That's the entire design of CosNFX — and there's honest physics underneath it.
| Link | Typical delay | Does it read? |
|---|---|---|
| Gesture → watch sensor | <50 ms | Immediate — the app plays locally |
| Bluetooth speaker hop | 150–300 ms | Noticeable on sharp cues; plan for it |
| Phone-controlled props | 100–400 ms | Use for glows and ambience, not hits |
That's why CosNFX plays its sounds straight from the watch's speaker first: the hit lands with the gesture. Props join for the layers (glow, fog, flash) where a couple hundred milliseconds reads as drama, not error.
Small watch speakers favor mids and highs: design cues around transients (clash, whoosh, chirp) rather than sub-bass rumble. And in a scene, the cue should land on the end of a motion — the strike, not the wind-up — or it reads as early. The app's gesture calibration is essentially teaching it your end-of-motion.