Number theoretic transforms to implement fast digital convolution

Transforms using number theoretic concepts are developed as a method for fast and error-free calculation of finite digital convolution. The transforms are defined on finite fields and rings of integers with arithmetic carried out modulo an integer and it is shown that under certain conditions this gives the same results as conventional digital convolution. Because of these characteristics they are ideally suited to digital computation by taking into account quantization of amplitude as well as time in their definition. When the modulus is chosen as a Fermat number a transform results that requires only on the order of N log N additions and word shifts but no multiplications. In addition to being efficient, they have no roundoff error and do not require storage of basis functions. There is a restriction on sequence length imposed by word length and a problem with overflow but methods for overcoming these are presented. Results of an implementation on an IBM 370/155 are presented and compared with the fast Fourier transform showing a substantial improvement in efficiency and accuracy. Variations on the basic number theoretic transforms are also presented.

[1]  G. Hardy.  The Theory of Numbers , 1922, Nature.

[2]  J. Tukey,et al.  An algorithm for the machine calculation of complex Fourier series , 1965 .

[3]  Thomas G. Stockham,et al.  High-speed convolution and correlation , 1966, AFIPS '66 (Spring).

[4]  Donald Ervin Knuth,et al.  The Art of Computer Programming , 1968 .

[5]  Donald Ervin Knuth,et al.  The Art of Computer Programming, Volume II: Seminumerical Algorithms , 1970 .

[6]  R. Singleton.  An algorithm for computing the mixed radix fast Fourier transform , 1969 .

[7]  Leland B. Jackson,et al.  On the interaction of roundoff noise and dynamic range in digital filters , 1970, Bell Syst. Tech. J..

[8]  Peter J. Nicholson,et al.  Algebraic Theory of Finite Fourier Transforms , 1971, Journal of computer and system sciences (Print).

[9]  I. J. Good,et al.  The Relationship Between Two Fast Fourier Transforms , 1971, IEEE Transactions on Computers.

[10]  J. Pollard,et al.  The fast Fourier transform in a finite field , 1971 .

[11]  A. Oppenheim,et al.  Effects of finite register length in digital filtering and the fast Fourier transform , 1972 .

[12]  Charles M. Rader,et al.  Discrete Convolutions via Mersenne Transrorms , 1972, IEEE Transactions on Computers.

[13]  C. Burrus,et al.  Block realization of digital filters , 1972 .

[14]  C. Rader.  A note on exact discrete Fourier transforms , 1973 .

[15]  C. Burrus,et al.  Fast Convolution using fermat number transforms with applications to digital filtering , 1974 .

[16]  Ramesh Chandra Agarwal,et al.  ON REALIZATION OF DIGITAL FILTERS , 1974 .

[17]  C. Burrus,et al.  Fast one-dimensional digital convolution by multidimensional techniques , 1974 .