Active listening in an acoustically crowded environment requires both maintaining and switching attention between auditory objects. General behavioral costs associated with task-switching are well documented in the cognitive literature. Yet there are many unanswered questions about the cost associated with switching between auditory objects relative to the type of feature-based attention employed. For example, do listeners use the same strategy to switch attention between locations as would be used to switch attention between speakers with different pitches? Is it equally difficult to switch attention within the same feature (e.g., attending to the left then the right talker) compared to switching across features (e.g., attending to the left talker then the one with the higher pitch)? Are the neural substrates recruited differentially depending on whether a listener is attending to spatial or non-spatial features of the auditory stimuli? In this talk, evidence from human psychophysical and physiological studies will be presented suggesting that auditory attention switching is not a monolithic construct. The interaction between top-down attention and different degrees of auditory stream degradation will also be discussed.