Video corpus of team meetings in immersive virtual reality

This corpus was collected as a part of the research project “AIDA: Action and Intersubjectivity in the Digital Age” (funded by the Research Council of Finland, 2023–2027). The data in the corpus consists of recorded team meetings held in a social virtual reality (VR) environment. In each recorded meeting, there are two to five participants present. All of the participants were familiar with one another prior to the data collection. The language of the data is Finnish. The data collection began in November 2023 and ended in May 2024. The team meetings were held in a commercial VR application called Immersed. Immersed has been designed specifically for co-working between remote participants. In Immersed, people can meet up with each other in public co-working spaces, or private collaboration rooms. Immersed allows people to connect their computer screens to the VR environment, and to share one of the screens with their co-workers. In this data corpus, the participants mostly meet in private collaboration rooms, and share or attempt to share their screens with each other in every meeting. The participants used Meta Quest 3 head-mounted displays (HMD) and hand-held controllers to access the virtual environment and interact with other users. In addition to using controllers, Meta Quest 3 supports hand-tracking, which means that the users can interact with the user interface (UI) without controllers, only with using their hands. The hand-tracking function also depicts the users’ hand movements to other users in Immersed. The participants were recorded from two different perspectives. First, each participant used the HMD to record their first-person perspective of the VR environment. Second, in most cases, the participants’ physical environments were recorded with GoPro cameras. The sounds were recorded with the HMDs’ built-in microphones. Currently (April 2024), the corpus includes 11 recorded team meetings. The duration of each the meeting ranges approximately from 30 minutes to 2 hours. The whole corpus amounts to approximately 14 hours of video recordings (approximately 690 GBs).

本语料库系研究项目「AIDA：数字时代的行动与主体间性」（由芬兰研究理事会资助，2023—2027年）的配套采集数据。该语料库收录了社交虚拟现实（VR）环境下录制的团队会议数据：每场会议共有2至5名参会者，且所有参会者在数据采集前便已相互熟识。本数据的语言为芬兰语，采集工作始于2023年11月，于2024年5月结束。本次团队会议依托商用VR应用Immersed开展，该应用专为远程参与者的协同办公场景设计。在Immersed平台中，用户可在公共协同办公空间或私密协作会议室中会面；平台支持用户将电脑屏幕接入VR环境，并可与协同工作者共享其中一块屏幕。本语料库中的参会者大多选择在私密协作会议室开展会议，且每场会议中均会共享或尝试共享彼此的屏幕。参会者通过Meta Quest 3头戴式显示器（HMD）与手持控制器接入虚拟环境，并与其他用户交互。除手持控制器外，Meta Quest 3还支持手部追踪功能：用户无需借助控制器，仅靠双手即可与用户界面（UI）完成交互，且该功能可将用户的手部动作同步展示给Immersed平台中的其他参会者。本次录制采用双视角方案：其一，每位参会者通过自身佩戴的HMD录制VR环境中的第一人称视角画面；其二，多数情况下，参会者的实体环境由GoPro相机录制。音频数据则通过HMD的内置麦克风采集。截至2024年4月，本语料库共包含11场录制完成的团队会议，单场会议时长介于30分钟至2小时之间，整份语料库的视频录制总时长约14小时，数据总量约690GB。