Forensic restoration of distorted audio

I have a sample from an Aphex Twin song (Milkman), it's speech that has been quite messed up.  It looks like it has the frequency patterns you get from 8 bit audio with no dither, or using the low quality settings when pitch-shifting very specific noises

Anyhow, I remember seeing that SteveG is a master of forensic audio (who said flattery won't get you anywhere?) and wondered if you (or anyone else) knew how to decipher what is actually being said..  Not for any particular reason, just something I'm curious about

Here is the sample of audio (1.23mb)
and here is the CEP filter for loading/saving APE audio (lossless compression) (0.07mb)

Thanks, happy playing

Both links do not work.

I took the files down, there were 2 replies to this thread which gave me the answer

If I remember, the sample had been modulated at about 790Hz, so when I demodulated it against a tone you got out something more listenable..   SteveG thought it was a subtle way to introduce some swearing into the forum..