Fig. 1From: A large TV dataset for speech and music activity detectionGenre distribution in the TVSM datasetBack to article page