A comprehensive investigation of the detection of applause sounds in audio signals is presented. It focuses on the processing of single-channel recordings in real time with low latency. Of particular concern are the intensity of the applause within the sound mixture and the influence of interfering sounds on the recognition performance, which is investigated experimentally. Various feature sets, feature processings, and classification methods are compared. Low-pass filtering of the feature time series leads to the concept of sigma features and yields further improvement of the detection result.