Explanation is needed for der_with_vad.py

[kylemin]

There are so many magic numbers in der_with_vad.py.
I'd like to know how to make the final output json file from all the results (vad, wearer's voice detection, audio matching, etc).
Could you provide more detailed explanation for this file?