Gabe proposed (and I agree) that we round our decibel readings down to integers (i.e. bytes). We don't know the true accuracy of our current floating-point readings, the microphones and software probably aren't more accurate than +/- 1 dB, and we can get meaningful enough data out of integer decibel readings.
This rounding should probably occur during the JSON conversion so that our dB averaging can continue to benefit from floating-point precision. A floor() and clamp to reasonable limits would probably be enough.
Gabe proposed (and I agree) that we round our decibel readings down to integers (i.e. bytes). We don't know the true accuracy of our current floating-point readings, the microphones and software probably aren't more accurate than +/- 1 dB, and we can get meaningful enough data out of integer decibel readings.
This rounding should probably occur during the JSON conversion so that our dB averaging can continue to benefit from floating-point precision. A floor() and clamp to reasonable limits would probably be enough.