Read error correction in software

[Chuck Guzis]

On 25 Jul 2007 at 20:43, dwight elvey wrote:


> If you take the calculated crc with the recorded crc, it can be
> played backwards. When you get close, the end numbers in your
> CRC back calculator will be zeros and the bits in the middle will
> be the error mask.

I'd thought of that, but most of the error sectors have more than a a 
single error (the content is mostly ASCII, so single bit errors are 
easy to correct by inspection).  Multiple MFM errors, what with bit 
and clock/data phase shifting etc. just don't seem to be amenable to 
that sort of approach.  Maybe I'm missing something...

Cheers,
Chuck