Compare cpm counter out with cpm radlog in

I am also comparing the CPMs measured by Geiger counter and those counted by RadLog.
I connected RadLog with audio microphone jack input
I think it's normale differecies at the same time, but macro differencies on an average in the same day it's a bit suspicious, correct?
in the same day on the one hand there are 89 cpm on the software there are over 120 cpm..
where am i wrong?
what did I forget to set in the software?
 

the spikes are even more suspicious
here is an example:
 

I had taken the signal from the piezoelectric capsule. I didn't realize the signal was so dirty and jagged.

I have now updated the peak recognition configurations in the "Audio" section of RadLog. So, now the software is counting the pulses more like the geiger counter source. In the meantime I will try to fix the signal upstream.