Datasets > Modality > Audio > CHiME-Home

CHiME-Home is a dataset for sound source recognition in a domestic environment. It uses around 6.8 hours of domestic environment audio recordings. The recordings were obtained from the CHiME projects – computational hearing in multisource environments – where recording equipment was positioned inside an English Victorian semi-detached house. The recordings were selected from 22 sessions totalling 19.5 hours, with each session made between 7:30 in the morning and 20:00 in the evening. In the considered recordings, the equipment was placed in the lounge (sitting room) near the door opening onto a hallway, with the hallway opening onto a kitchen with no door. With the lounge door typically open, prominent sounds thus may originate from sources both in the lounge and kitchen.

The choice of permitted labels was motivated by the sources present in the considered acoustic environment: Human speakers (c,m,f); human activity (p); television (v); household appliances (b). Further labels o,S,U respectively relate to any other identifiable sounds, silence, unidentifiable sounds. Labels S,U may respectively only be assigned in isolation. Annotators were acquired to assign at least one label to a chunk, thus annotators may either assign one or more labels from the set {c,m,f,v,p,b,o}, or may alternatively ‘flag’ the chunk using a single label from the set {S,U}.

Source: https://www.researchgate.net/publication/308732345_CHiME-Home_A_dataset_for_sound_source_recognition_in_a_domestic_environment