Audio file re-analysis occasionally leads to different results

#2
by MWeber - opened

I encountered a strange situation: yesterday I analyzed an audio .wav file (5 sec ultrasound recording using Phil Atkin's PiPistrelle hardware at a maximum of 384 kHz sample rate) which gave no hit (no bat detected). Today, I re-analyzed that same file and it gave several species hits. Here is that file:

Not sure what could have caused the issue. Looks as if some data is carried over - maybe a RAM leak in the code?
Would be important to investigate and to correct as I am not sure whether this also affects the GitHub code and the corresponding scientific publication.

Michael.

Sign up or log in to comment