A method of encoding that uses standard codecs such as linear encoders and decoders for encoding and decoding data with different levels of robustness to errors is described. In one configuration, multiple encoders may be utilized, and one of the encoders may use a standard encoder such as a turbo code followed by a nonlinearity that creates an unequal distribution of ones and zeros in a binary representation of the code. In another configuration, a coder may be utilized that represents message outputs as “channels” that create state transitions (or symbol errors) in a data forward error correction coder.