This post will be quick and to the point. We’ll be using SoX to bring high frequency audio, commonly referred to as ultrasound, down to a range that can be heard.

First thing I did to validate the process was generate an ultrasonic data file using Audacity. The problem with the standard 44.1k sampling rate is ultrasound frequencies will be dropped. I created my data files with a sampling rate of 192k.

play "/home/local/Desktop/Ultrasound/Audacity192k_33khz.wav"

Playing audio from this sample is outside human hearing, nothing appears to be playing. I’ll use this command to bring the pitch down so it can be heard.

play "/home/local/Desktop/Ultrasound/Audacity192k_33khz.wav" pitch -q -3000

Now I’ll save a new file with the pitch shift. Here is the command.

sox "/home/local/Desktop/Ultrasound/Audacity192k_33khz.wav" "/home/local/Desktop/Ultrasound/Audacity192k_4khz.wav" pitch -q -3000

When doing a spectrogram, the scaling is too high because of the initial 192k sampling rate. I dropped it down with this command.

sox "/home/local/Desktop/Ultrasound/Audacity192k_4khz.wav" -r 22500 "/home/local/Desktop/Ultrasound/Audacity22k_4khz.wav"

The following video has crossfade that goes between normal and pitch reduction. The high pitch of engines and breaks can be heard easier with the frequency dropped down. There are occasional beeps that can also be heard, I suspect these are proximity sensors that newer cars have.

Not much to this topic. Again, SoX shows that it can operate in ultrasonic ranges. I hope you enjoyed this post.