We investigate segmental and prosodic realizations across two varieties of a speaker (Zurich German and Swiss Standard German). Ten native speakers from the canton of Zurich read alternative questions of the type ‘Do you want or X?’ in their dialect (Zurich German) and in their standard variety (Swiss Standard German). Ten Standard German speakers from Germany read the standard stimuli as control. We analyzed word-initial stops and pitch accent realization of prenuclear rising accents on disyllabic, trochaic target words (e.g., lenis: backen ‘to bake’ vs. fortis: packen ‘to pack’). The results showed that voice-onset-time (VOT) for fortis stops was the longest in Standard German, followed by Swiss Standard German, which in turn had longer VOT than Zurich German; VOTs for lenis stops did not differ across varieties. Swiss Standard German fortis stops were numerically closer to Standard German controls than Zurich German stops were. Intonation was modelled using general additive mixed models. The results showed that, within-participants, prenuclear accentual rises in Zurich German had a smaller f0 range than in Swiss Standard German. In comparison to Standard German Swiss Standard German was very similar in f0, but Zurich German differed for most parts of the target word.