complex_median may merely be minimum intensity

[markromanmiller]

Test whether the complex_median method may be approximated by the minimum magnitude method, which would mean that if there's a recording with muted sound, the sound would be significantly affected.

[markromanmiller]

Solution: Normalize all the recordings (zero mean, same variance). This needs to be added into the code.

[timthejiang]

I created a separate branch tim.

What I've done so far:

• Handle frequencies of silence; commits #1 and #2
• Normalize the recordings. I'm not sure if I did it correctly, as the output seemed the same with input being normal and silent recordings. I looked up normalizing on StackOverflow; commits #3

@MrMallIronmaker , I would appreciate your feedback and suggestions.