[cdp] more re DISTORT REPEAT

[Archer Endrich]

Hi Peter,

My colleague was very quick off the mark.  He writes:

"ah! The problem (if it can be called that) is that the distort 
processes depend uniquely on the exact contents of the input file, in 
finding "wavecycles" - pairs of zero-crossing points. In the case of the 
"fails" file, it has a negative DC offset such that a whole chunk of the 
data is all below zero, so no crossings found. That is what the message 
signifies. A simple inspection with Audacity will reveal this. I think I 
may have documented something about this years ago. It will be necessary 
to pass such a file through a DC removal filter. I don't think CDP 
actually has a "dcremove" process as such (there is a handy Csound 
opcode for it though!), but any high-pass filter with a low centre freq 
such as 20Hz (higher is better if the file is not LF-heavy) should do 
the trick. I think I also recommended a low-pass filter as well to 
remove high frequencies which generate loads of tiny HF wavecycles which 
are likely not to give expected or desired results."

I didn't remember this -- thanks for your query.

Cheers,

Archer - CDP