The goal of this study is to provide a multidimensional description of rhythmic structure of Polish utterances. For this purpose a time-delay approach proposed in [1] is applied and results of qualitative and quantitative analyses based on time-delay plots are compared with results obtained with selected rhythm metrics. The study shows that description that relies on a combination of rhythmic scores is inconclusive and difficult to interpret, because it does not account for rhythmic structuring nor grouping. The time-delay approach, on the contrary, appears to be very efficient in exploring short-time and long-term timing variability that determines Polish speech rhythm.