Listeners can judge the overall loudness of time-varying sounds quite easily, i.e., assign a single value that represents the loudness of the entire sound. This holds even if the duration is long and the judgment includes memory effects. Different metrics for calculating overall loudness have been developed. They agree that overall loudness is higher than the mean of loudness over time. Percentiles like the N5, the loudness being exceeded 5% of the time, are adopted by ISO 532-1. In the present study the concept of an energy mean known from level measurements (ISO 1996-1) was applied to the loudness domain. This equivalent continuous loudness level, LLP, was compared to the N5 using a set of real-world sounds that was orthogonal between the two metrics over a wide dynamic range of 30 phon. Cross-modality matching with line length was used in three experiments with a focus on either the overall judgment of loudness, continuous judgment while a sound was played, or both. The LLP showed considerably higher correlations with overall judgments than N5. Comparing continuous instantaneous judgment with calculated instantaneous loudness suggests that the participants might have focused on the sounds' prominent portions.