Question about wav2spec function in utils/audo.py

I have noticed that there is ```amp_to_db``` function in ```wav2spec```, which means that the input of model is db-scaled magnitude. I s this right? Since this is not mentioned in related paper.