The teachings presented herein improve the processing of individual signals of interest included in a received composite signal by computing combining weights and/or signal quality estimates for each signal of interest, e.g., for linear equalization, based on either shared or non-shared correlation estimates. As a non-limiting advantage, the use of shared correlation estimates reduces computational loading as compared to the processing load that would be needed for computing non-shared correlation estimates for all signals of interest. As a further non-limiting advantage, the conditional use of shared and non-shared correlation estimates provides for the use of non-shared correlation estimates where signal characteristic(s) of one or more of the signals of interest warrant such usage, e.g., for one or more high-rate signals of interest.