Find the Most Constant Minimum Difference in a 2D array

I have a 2-Dimensional Array where each column refers to a different threshold value. Each row in a given column contains difference (A-B) results between two arrays. What I need to do is find out which of these columns consistently produces the least amount of differences (in other words, A-B stays between C [min] and N[max] for the entirety of the column, but where the values of C and N are unknown).

This question accompanies this older question , however this older question is now irrelevant as the only part of the code I don't know how to figure out yet is this part.

Answer 1

[Matt J]

So you're trying to find which column has the minimum span semi-norm?

span=max(A,[],1)-min(A,[],1);

[deltaMin,loc]=min(span);

Answer 2

[Image Analyst]

How can C and N be unknown? I can tell you 35% of the elements in column 1 are between 5 and 10 if I know what 5 and 10 are. But if you're not going to tell me 5 and 10, then how am I supposed to know what your lower and upper range values are?

Why does C and N matter anyway? Why can't you just do

columnMeans = mean(abs(array2D), 1)
[leastDiff, colWithLeastDiff] = min(columnMeans)

to find the column that has, on average over all the rows, the least absolute difference?