Discrete-time modeling of transition-noise-dominant channels is considered which facilitates performance analysis of various sampled-data detection schemes. Based on the proposed channel description method in the presence of transition noise, expressions for signal-to-total-noise-power ratios associated with a few selected detection schemes are derived. Under the assumption of a Lorentzian step response and perfect equalization, comparison is made among different detectors based on the signal-to-total-noise ratio figure-of-merit evaluated as a function of linear density.