Default to mundane
Every credible investigator starts with 'what could explain this naturally?' before 'what if it's paranormal?' This isn't being a skeptic killjoy — it's the only way to ever produce evidence that's worth taking seriously. If you can't rule it out, that's when it gets interesting.
Common false positives
Dust orbs (close to the lens, lit by flash). Reflection ghosts (window glass, polished surfaces, screens). EMF spikes (wiring in walls, fridges, phones, cheap LED drivers). Audio 'whispers' (HVAC, distant conversation, breath on the mic). SLS skeletons (chairs, coat racks, plants — anything roughly humanoid).
Run the 'try to recreate it' test
If you got a high reading, try to reproduce it: walk the same path, run the appliance, open the window. If you can recreate the signal at will, you've found the cause. SpiritMesh logs these as Debunked, which actually raises the credibility of your other clips.
Confidence is a number, not a verdict
85% confidence does not mean ghost. It means the signal pattern matched what your sensors were primed to detect. Always pair confidence with context: baseline, environmental notes, and any reproduction attempts.