This paper presents a methodology for the study of audience perception of live performances, using a combined approach of post-hoc and real-time data. We conducted a study that queried audience enjoyment and their perception of error in digital musical instrument (DMI) performance. We collected quantitative and qualitative data from the participants via paper survey after each performance and at the end of the concert, and during the performances spectators were invited to indicate moments of enjoyment and incidences of error using a two-button mobile app interface. This produced 58 paired post-hoc and real-time data sets for analysis. We demonstrate that real-time indication of error does not translate to reported non-enjoyment and post-hoc and real-time data sets are not necessarily consistent for each participant. In conclusion we make the case for a combined approach to audience studies in live performance contexts.