Median filter isn’t going to help “Original.mp3”, the median filter taken over 3-sample points only gets rid of high-frequency crackle, your “Original.mp3” sample is all below 4kHz so doesn’t have any high frequency.
Your example needs heavy-duty noise-reduction.
There is a free plugin [Windows&Mac] called DtBlkFx which has a “contrast” effect , which is effectively dynamic noise reduction, a before-after example attached with contrast set at at 20%, 30%,40% & 50% …
The higher the contrast the more computery (synthesised) it sounds.
