Abstract Computing dissimilarity profiles between spectra within a chromatographic peak and comparing them to a base profile for a pure peak can be a sensitive method for detecting coelution. The power of this method is limited by noise and non-idealities. For high maximum absorbances, non-idealities prevail over noise. In current practice, either the maximum absorbances are limited, or the base profile is computed from injections of pure standards. The first approach leads to a loss in detection power, the second requires substantial extra work, since the standards have to be obtained and they have to be measured separately for each peak of interest in the chromatogram. Here, we show that an approximation to the base profile can be computed from realistic simulations of homogeneous data accounting for potential non-idealities. This is demonstrated using several graphical displays. As a result, higher maximum absorbances leading to better impurity detection can be used without extra experimentation.