A method for encoding and decoding digital data for storage in a holographic medium (12). Digital data, consisting of binary data (B.sub.i) or grey scale data (A.sub.i), is encoded in bit groups or digit groups (B.sub.k, A.sub.k) containing at least k=1 bits or digits, respectively, by assigning to each bit group (B.sub.k) one reference bit (B.sub.r) and to each digit group (A.sub.k) two reference digits (A.sub.r1, A.sub.r2), assigning the bits of group (B.sub.k) to information bits (B.sub.j), assigning the digits of group (A.sub.k) to information digits (A.sub.j), assigning the reference bit (B.sub.r) and information digits (B.sub.j) to a reference pixel (P.sub.r) and information pixels (P.sub.j) chosen from pixels (24) of a holographic signal modulator (18), and assigning the reference digits (A.sub.r1, A.sub.r2) and information digits (A.sub.j) to reference pixels (P.sub.r1, P.sub.r2) and information pixels (P.sub.j) chosen from pixels (24) of the holographic signal modulator (18). Storing all pixels (24) by illuminating the holographic signal modulator (18) with a signal beam (22) which impresses upon the signal beam (29) pixels (24), guiding the signal beam (22) to the holographic medium (12) and storing the pixels (24) by directing a reference beam (26) at the holographic medium (12) from a predetermined angle .beta. and at a predetermined part of the holographic medium (12). Decoding the encoded binary data (B.sub.i) and encoded grey scale data (A.sub.i) by aiming the reference beam (26) at the holographic medium (12) at the same angle .beta. and location as during storage, recovering pixels (24) and recovering from them using a decoding unit (38) reference bits (B.sub.r) and information bits (B.sub.j) for each bit group (B.sub.k), and recovering reference digits (A.sub.r1, A.sub.r2) and information digits (A.sub.j) for each digit group (A.sub.k). Reconstructing the original binary data (B.sub.i) by differentiating the binary values of bits (B.sub.r) and (B.sub.j). Reconstructing the grey scale data (A.sub.i) by differentiating the reference digits (A.sub.r1, A.sub.r2) to obtain a benchmark (D.sub.r) and applying a binary operation on the benchmark (D.sub.r) and each of the information digits (A.sub.j).