Immersive Virtual Environments combined with motion capture systems have been used as experimental set-ups for studying musicians' performances in different contexts. By exploring both acoustic and motion features, this study highlights the impact of the presence of a virtual audience on both the performance and the perception of authenticity and emotional intensity by the listeners. Gestures and sounds produced were impacted differently when musicians were asked to play with different expressive intents. The presence of an audience reduced the differences between those categories and makes them converging towards values associated with the projected expressive manner. On the listeners' side, only the congruent situation of projected expressive intent and the presence of an audience boosted the participants' ratings in both authenticity and emotional intensity. When comparing the different features values, stimuli recorded with an audience were associated with a bipartite distribution of the values for both authenticity and emotional intensity (either high or low) contradicting the three artificially created expressive categories. This study highlights the use of IVEs as a research tool as well as a training assistant for musician eager to learn how to cope with audience anxiety.