DrCoffey
commented
3 years ago Ya, the Dodotronic mics are notorious for terrible noise bands. Sometimes you can hunt down the noise if it is external (like a fluorescent light or monitor or whatever). But sometimes the mics just have strange electrical noise you can't remove.
I think you can preprocess the recordings with a notch filter to cut out a very thin audio band. I think you can do it in audacity or in Matlab. In Matlab the code is something like this (say it is a 45kHz noise band).
[x, fs ]=audioread('audiofile.wav');
wo = 45000/(fs/2);
bw = wo/35;
[b,a] = iirnotch(wo,bw);
y = filter(b,a,x);
audiowrite('filteredaudiofile.wav',y,fs);
https://www.mathworks.com/help/dsp/ref/iirnotch.html https://www.mathworks.com/help/matlab/ref/filter.html
In the long run I definately reccomend upgrading to something like the Pettersson M500-384 https://batsound.com/product/m500-384-usb-ultrasound-microphone/
Hi Dr. Coffey, I’ve been using DeepSqueak V3 for a few days, and it is has been very helpful. However, Ive encountered a slight issue with detecting calls; I have a few noise bands that DeepSqueak keeps detecting as calls. I got rid of the biggest noise band by adjusting the frequency threshold, but there are still a few smaller ones which interfere with the detection. I suspect these noise bands are a byproduct of the mic I use, the dodotronic ultramic. Is there any way to get around these noise bands without having to buy a higher quality mic?