Previous studies have revealed that, depending on the language, listeners can identify speakers’ dialects quite well. The role of segments and prosody in this task is largely unknown, however. In a between-subjects design, we tested a total of 30 listeners in two conditions: in the unmorphed condition, listeners heard original sentences from two Swiss German dialects; in the duration morphed condition, listeners heard the same material, but with syllable durations exchanged between the two dialects. In a two-alternative forced choice design, subjects judged the regional origin of the stimuli heard. Results revealed near perfect identification performance for both conditions, thus underlining the overriding dominance of segmental cues in dialect identification tasks. The findings reported are pertinent to forensic phonetics, enhancing the diagnostic power of naïve and expert listeners’ claims about suspect speakers’ voices.