A gallery added short audio descriptions to several displays. The recordings began by directing attention to a visible detail before introducing an interpretation. Visitors spent longer at those displays, which initially pleased the curators. Yet a follow-up conversation revealed that some visitors remembered the narrator's dramatic voice without being able to identify the detail discussed. The gallery revised these recordings, leaving pauses in which visitors could inspect the object themselves. Written labels remained available for people who preferred reading or could not use the audio. The aim was not to replace every form of explanation with sound. It was to connect an explanation to an act of observation. More time spent at a display could indicate interest, confusion, or both; the curators needed another kind of evidence before treating that time as proof of understanding.