A microphone in a small room records a coloured sound: the room's resonances boost some frequencies and the mic adds its own curve. KAMI takes a REW frequency-response measurement, finds the peaks and nulls, and writes an EQ correction profile that flattens the response.
What I built
A database that stores every raw measurement by content hash, so each export traces back to its source.
Two correction algorithms (wide-peak cut, peak/null mirror) with gains limited to −6 to +6 dB.
A local web interface for comparing rooms and microphones.
Verification
20 unit tests passing.
Reproduces the earlier script outputs byte for byte, including rounding behaviour.