Honestly: don't know.

My experiments never had a dependency on linearity.

To detect something, you at least need response that isn’t effectively zero at the relevant frequency. And bonus points for not having ludicrous amounts of jitter due to polling a sensor over a bus like I2C that is really not intended to collect high frequency, equally spaced samples.

It "looked good" for me up to 100Hz after Nyquist in the spectrum but didn't pull any precise data. Just created a 50Hz tone using my speakers which showed up in the spectrum afterwards as expected as basic verification

Took some spectras from heat pumps, airplanes and a helicopter.

So the sensor isn't blind within this range.

Huh, impressive.

A long time ago I used a hot wire anemometer to measure flow up to several hundred Hz, but that was an expensive instrument being read by an expensive National Instruments acquisition platform. This is a cheap, self-contained device. Go Sensirion.

I wonder whether the frequency limit ends up being set by the ADC or the physical sensor. I’m also curious how the Sensirion sensor detects the sign of the pressure difference.