The complexity of a CELP algorithm is mainly concentrated in the codebook search procedure. Many studies show that the algebraic approach for codebook structures considerably reduces the computational load with a slight degradation compared to stochastic or trained codebook performances. We propose a novel search algorithm based on a Hamming single-error-correcting -code (ECC) structure. The optimum codevector is directly circumscribed using the error correction ability of the ECC. This algorithm requires only 0.7 MIPS which represents a complexity reduction by a factor of 3 or 4 compared to search algorithms based on current algebraic codebooks.