Poster presented at the Speech in Noise Workshop 2020, Toulouse, France. 9-10 January 2020 Abstract: In everyday life, listeners are often confronted with situations in which multiple talkers speak simultaneously. Those so-called cocktail-party situations can be static, that is, the target talker always remains the same, or they can be dynamic, meaning the target talker changes in an unpredictable manner. Previous studies have shown that, in dynamic situations, the listener’s speech intelligibility often decreases after a transition from one target talker to the other. Afterwards it takes some time until the speech intelligibility is back to its initial state again. This study examined which factors contribute to the drop in speech intelligibility after a transition. To this end, a setup involving three competing talkers uttering matrix sentences was used (cf. Meister et al., this conference). The talkers differed in terms of their voices (medium male, deep female and high female) and spatial positions (-60°, 0° and +60° azimuth angle). The target talker was unknown to the listeners in advance and was indicated by the first word of the sentence. Depending on the condition, target talkers changed either after each trial (transition probability = 100%) or only in 20% of the trials. Two different transition types were used: 1) The talkers either remained at the same positions and the target voice changed; 2) the talkers switched positions but the target voice remained the same. Fourteen younger normal-hearing and 18 older normal-hearing listeners participated in the study.This poster will present an analysis of the influence of different factors on speech intelligibility after transitions. These factors include the target talker position, the transition probability and the individual hearing thresholds. Furthermore, we will address the question of how fast listeners were able to switch attention from one talker to the other.Funding: Supported by Deutsche Forschungsgemeinschaft (ME2751/3-1).