On February 24, 2009 at 13:22, Healey3000 said...
They are arranged as three groups of 16, with longish wait states between them.
That is absolutely NOT Kaseikyo format.
I then pulled down a couple of ccf files for the 2802 off the files section and aaronmac's formats match mine so far (I've only looked at three or four keys). Interestingly enough, Joel's file has some strange ones in there that match no format at all (that I can tell).
I'll try later to figure out which CCF's you mean, download them and use DecodeCCF to verify, but I'm confident I'll find Denon's version of Sharp protocol and the strange ones are probably bad learns (surprisingly many signals in posted CCF files are garbage that would never work).