Voice control is an increasingly common feature of digital games, but the experience of playing with voice control is often hampered by feelings of embarrassment and dissonance. Past research has recognised these tensions, but has not offered a general model of how they arise and how players respond to them. In this study, we use Erving Goffman’s frame analysis [16], as adapted to the study of games by Conway and Trevillian [9], to understand the social experience of playing games by voice. Based on 24 interviews with participants who played voice-controlled games in a social setting, we put forward a frame analytic model of gameplay as a social event, along with seven themes that describe how voice interaction enhances or disrupts the player experience. Our results demonstrate the utility of frame analysis for understanding social dissonance in voice interaction gameplay, and point to practical considerations for designers to improve engagement with voice-controlled games.