Voices+Voids: Reclaiming and Transcoding Our Data as Performance