Summary: We propose a storage scheme for a string \(S[1,n]\), drawn from an alphabet \(\Sigma\), that requires space close to the \(k\)-th order empirical entropy of \(S\), and allows one to retrieve any substring of \(S\) of length \(\ell\) in optimal \(O(1+ \frac{\ell}{\log_{|\Sigma|}n})\) time. This matches the best known bounds via the use of binary encodings and tables only. We also apply our storage scheme to the Burrows-Wheeler Transform and achieve a space bound which depends on both the \(k\)-th order entropy of \(S\) and the \(k\)-th order entropy of its BW-transformed string \(\text{bwt}(S)\).