GaMMA Corpus: Polyadic Conversations in Danish with Gaze, Speech, and Motion Data in Quiet and Adverse Speech Conditions

The GaMMA (Gaze, Motion, and Multi-talker Audio) corpus captures the behavior of polyadic conversations among native Danish speakers under both normal and cocktail party conditions. Eleven groups of four normal-hearing participants (44 unique individuals) are recorded while engaged in natural and spontaneous interactions. All conversations were conducted without conversational tasks. Each group was intentionally composed of participants with prior intragroup and interpersonal relations. Gaze and motion data were collected using an optical tracking system (Vicon) and eye-tracking glasses (Tobii), while speech was recorded via omnidirectional head-worn microphones and binaural hearing aid microphones with low occlusion. Calibrations were conducted before trials and compensation filters were created to account for differences in microphone placements. Processed versions of the audio signals, with background noise attenuated and crosstalk removed, were used to compute speech activity for all participants. The corpus, including both raw and processed gaze and audio data, as well as filters, calibration signals, and speech activity output, is included in this dataset.The Data descriptor for the corpus titled "The GaMMA corpus of Danish polyadic conversations with gaze speech and motion data in quiet and noise" can be found via this DOI: (will be added once journal finishes final review).Full annotations for the corpus are available at: https://doi.org/10.5061/dryad.r7sqv9snc (in review, DOI not active yet)and described in this article: https://aclanthology.org/2025.sigdial-1.19/