N-digit, radix-alpha bases are proposed by VLSI implementation of redu
ndant arithmetic, mod m, where [alpha(N)](m) = +/-1, [alpha(i)](m) not
equal +/-1, for 0 < j < N and m is prime. These bases simplify arithm
etic overflow and are well suited to redundant arithmetic. The represe
ntations provide competitive, multiplierless T-point Number Theoretic
Transforms, mod m, where Tl N or Tl 2N.