“…Indeed, especially in multimodal communication research on humans, original data that support one's analyses are often not openly shared, because they often consist of audiovisual recordings of identifiable people. There is, however, an increasing number of tools that allow to partially mask the identities from video and audio automatically, while still extracting non-identifiable information that can support analyses, such as facial, hand, and body pose information [82,133,156]. It is important to note that these tools do not count as anonymization tools, because either the transformed sound is still re-transformable to its original (thereby allowing identification in principle) or is still present next to a bodily mask.…”