Whistle detection ⁑ Dercuano

submited by

Style Pass

2022-05-13 06:00:05

I can whistle from about 600Hz up to about 1600Hz. There is often substantial unintentional amplitude modulation on my whistle around 25–100 Hz. The vibrations can’t ramp up or down by 10dB in less than about 8 cycles, even when I’m releasing a tongue stop, and more typically take 20 or 30 cycles, suggesting a Q factor of somewhere around 10–20. Second, third, and fourth harmonics are detectable, but weak; even the second harmonic is 52dB down from the fundamental. I can easily ramp up from 800Hz to 1600Hz in 1.3 seconds. Doing the same ramp in 100 milliseconds is feasible with more effort.

(Semitone resolution is a Q factor of about 17, so using a linear filter with a lower Q factor would fuzz out frequency selectivity too much to be useful.)

A particular 70-millisecond segment of low whistle had a frequency peak at 615 Hz at -29.3 dB, falling off to -35.6 dB at 610 Hz and -38 dB at 628 Hz. This suggests a Q factor of around 60 during the whistle itself. (Maybe even a bit higher.)