Disk Nibbling

[Zac67]

msh5150 wrote:

I've been wondering what exactly "nibble copying" is and how it works.

The difference is that you don't use the standard MFM->data de- and encoding but either non at all (see below) or custom routines.

It seems to me the biggest problems are finding the tracksync mark (especially if it's not $4489) and working out how long the track is?  :-o

That's the art of coping with protection schemes. If you can't analyse several 'identical' disks, you're much left to guessing. But the choice of sync words in MFM is limited.

As long as those problems are solved there is no reason to decode the MFM?

If you can't find out the right sync mark, chances are high that you mess up the track (some bits are always lost). Furthermore, if you don't de- and reencode MFM (or whatever is used!) minor read errors will get duplicated, too and add up from generation to generation.
Since Amy's doing channel en/decoding in software, you're not limited to MFM (like on Wintel), but can also do GCR or something completely different! As long as Paula can sync to the data stream there's no limit. ;-)

[Zac67]

That's what I was going to say: this is where it's getting illegal. ;-) I understood the question more like: how do I write a universal copy algorithm and not how do I crack this game...

[Zac67]

Well, I guess your experience grows with every protection scheme you analyse. ;-)

Of course there are methods that a standard drive can only read but not write, like long tracks (=slightly higher bit rate) or weak bits (=bits that may be read as 0 or 1 by chance), but fortunately these were expensive to produce and therefore quite rare. :-P

It's a shame there isn't a good Wikipedia article on MFM or floppy disk formats (little technical detail for my taste), but there's probably plenty resources around this topic on the net somewhere.