I’m analyzing sound recordings that include baseline noise, short impact sounds, and skin friction sounds in random order. In the FFT analysis in Audacity, I noticed a sharp drop in the frequency range between 90kHz and 95kHz for both the skin friction and background noise, but not for the impact sound. Could this drop be due to how Audacity processes signals in FFT, and if so, why wouldn’t it affect all types of sounds similarly?
i am no expert at all, but this looks like a normal limitation of some kind, hardware or software. the recovered part is probably an artifact.
it differs probably because the cracking thing is generated, in a different way.
i have a similar limitation at around 42-48KHz on my audio recordings. lowering the project rate(Hz) is lowering the cutoff, increasing it is increasing it to a point.
…btw, is this fft a plug-in or a new feature, as i dont have it?
The dip at ~90kHz on your spectrum is not due to Audacity.
Below I generated white noise at a sample rate of 192kHz, no dip …
Audio electronics are designed to work within human hearing range: nominally 20Hz-20kHz.
Audio electronics are designed to roll-off beyond 20kHz.
Even electronics designed to record ultrasound will have a frequency upper limit.