ignore chars with GNU diff

[Gary Haskins]

I am diff'ing two ascii text files with known differences at known
offsets in each record.  I want to find records with other differences
while ignoring the known differences.  I am looking for a diff option
which allows me to ignore chars 'at offset for length' or 'from begin
offset to end offset'.  I Googled and scoured the manual but cannot
find such an option.  I see the options for ignoring case differences
and lines which match a regexp, but these are not what I want (at
least I don't think so).  There are ways I can twiddle the data to
remove the known differences, but I was hoping it may be as simple as
just specifying a range of character offsets to ignore.  Have I missed
something in my search?  Does anyone know if GNU diff support this
kind of option?

OS is Suse and diff version is:
diff (GNU diffutils) 2.7.2
Copyright 1988, 89, 92, 93, 94, 96, 1998 Free Software Foundation,
Inc.

Tks,

Garyh