Matching Huffam quants

We encode separately from t or f(s) = t.  In the matching process we are finding the minimum  deviation pairs, not matching them by value. A spot on the Markov tree selects the relative M resolution, the first value in 1,x,y limits the error deviation.  It forces a finite deviation operator. ost solutions back to the original are about finding a good vertical from the placement of those spin markers.