Outputs from at least one pseudo-random source are used to encode hidden value. The hidden value is encoded using index based quantities, for example, based on numerically ordering a sequence of outputs from pseudo-random source(s). In some examples, the numerical ordering of re-generated device-specific quantities is used to re-generate the hidden value, without necessarily requiring additional error correction mechanisms. Information leak may be reduced by constructing system whose “syndrome” helper bits are random, as measured, for example, by NIST's Statistical Tests for Randomness In some examples, index based coding provides coding gain that exponentially reduces total error correction code complexity, resulting in efficiently realizable PRS-based key generation systems. In some examples, index based coding allows noisy PRS to be robust across conditions where conventional error correction code cannot error correct.