Simplification of the Computation of Real Hadamard Transfors

The computation of a real Hadamard (or Walsh) transform can be simplified by first computing a pseudotransform, employing an easily derived auxiliary matrix each of whose elements is either l or 0 (instead of 1 or -1), and then applyimg se very simple linear corrections to obtain the; desired transform. The auxiliary matrix is not orthogonal, so its square, unlike that of the Hadamard (or Walsh) matrix, is, not a scalar matrix. However, sibject to one easilyg satisfied restrietion, it has an inverse, Which, within a scalar. multiplier, differs from ihgt of the Hadamard (or Wtalh) matrix of the same order in only a single element.