Supporting collaborative transcription of recorded speech with a 3D game interface

The amount of speech data available on-line and in institutional repositories, including recordings of lectures, "podcasts", news broadcasts etc, has increased greatly in the past few years. Effective access to such data demands transcription. While current automatic speech recognition technology can help with this task, results of automatic transcription alone are often unsatisfactory. Recently, approaches which combine automatic speech recognition and collaborative transcription have been proposed in which geographically distributed users edit and correct automatically generated transcripts. These approaches, however, are based on traditional text-editor interfaces which provide little satisfaction to the users who perform these time-consuming tasks, most often on a voluntarily basis. We present a 3D "transcription game" interface which aims at improving the user experience of the transcription task and, ultimately, creating an extra incentive for users to engage in a process of collaborative transcription in the first place.