Mic DSP must be off — the page asks for it and reports what the browser actually gave.
You know how many blades your prop has — forcing it removes the one genuinely ambiguous step and makes every reading exact. Leave it on auto to let the rev-locked lines argue, e.g. on a toroidal where "blade count" is not well defined.
A brushless motor cannot exceed kV × volts. Setting this range from what you actually have excludes impossible answers on physics instead of arguing with them on spectra — it is the single strongest thing you can tell the tool. set max from kV × V
A prop radiates at the blade-passing frequency, BPF = rev/s × blades, plus harmonics. BPF alone cannot tell you the blade count, and guessing wrong scales the answer by 1.5×. So the blade count is resolved against two lines that depend on rev/s directly: the once-per-rev imbalance line, and commutation at rev/s × pole pairs. Only the right count puts both on real peaks. Candidate fundamentals are scored on their whole harmonic comb, not raw loudness — the 2nd harmonic is often stronger than the 1st and reads as double the RPM.
⚠ Record off-axis, out of the slipstream. Wind noise on the mic buries the tone.
Start the mic, or drop a file.