Homework 4 solutions, CS420, Fall 2011

1. Write a Java, C, or C++ program ... 2. Prove that when you take two polynomials of length n then add the two DFTs as vectors, you get the same result as you do when you add the polynomials together and then take the DFT of the result. Think about how we expressed the DFT in terms of a product of a matrix and a column vector. A column vector is just the special case of a matrix that has only one column. So use matrix algebra to prove it. Recall that matrix algebra must use only the algebraic laws of a ring. Solution: Let V be the DFT matrix. If x is the sequence of coecients for the rst polynomial, suitably padded with 0s, and y is the sequence of coecients for the second vector, suitably padded with 0s, then the sum of the the two DFTs are V x + V y. Since matrices on real or complex numbers form a ring, multiplication distributes over addition so this is the same as V (x + y). The latter is the result of adding the polynomials before taking the DFT. 3. Prove that if r is a real number and a is a vector, then the DFT of ra is r times the DFT of a.
n1 Solution: Let V be the DFT matrix. The j th element of the DFT of ra is i=0 xi (rai ), where xi is Vji (rai ), where ai is the ith element of a. By the distributive propery, this n1 is r i=0 Vji ai , which is r times the j th element of the DFT of ai Since j is arbitrary, every element of the DFT of ra is the same as the corresponding element of r times the DFT of a.

4. What is the DFT of (0, 1, 0, 0, 0, 0, 0, 0)? Solution:This is just the elements of the second row of the DFT matrix, namely, 7 1 0 (8 , 8 , . . . , 8 ). 5. What is the DFT of (1, 1, 1, 1, 1, 1, 1, 1)? There is a long-winded way to get this, and an easy way to get it if you remember some facts we used in showing that the DFT matrix has an inverse. Solution:The ith element of the result is just the sum of elements in the ith row of the DFT matrix. The elements in a row cancel, unless its the rst row. The result is (8, 0, 0, 0, 0, 0, 0, 0). 6. What is the DFT of (1, 0, 1, 1, 1, 1, 1, 1)? There is a long-winded way to get this also, but there is an easy way that makes use of the last four things you proved, above. Solution:This is just (1, 1, 1, 1, 1, 1, 1, 1) + (1)(0, 1, 0, 0, 0, 0, 0, 0). By the solutions to to Problems 2 and Problem 3, its DFT is the just the dierence of the DFTs of (1, 1, 1, 1, 1, 1, 1, 1) and (1)(0, 1, 0, 0, 0, 0, 0, 0). By Problems 4 and 5, this 7 2 1 7 1 0 is (8, 0, 0, 0, 0, 0, 0, 0) (8 , 8 , . . . , 8 ) = (7, 8 , 8 , . . . , 8 ).