Since being excavated in 2003, the skull of LB1 (the holotype of Homo floresiensis) has been given many faces, though the details regarding how each was accomplished are typically few. Here we detail our application of known, and verified, relationships between the skull and soft tissues of anatomically modern humans to produce an evidence-based facial approximation of LB1. We then compare our results to nine pre-existing LB1 faces using geometric morphometrics. These analyses suggest our facial approximation differs in proportional facial width, upper lip height and nasal morphology. Some of these differences are likely due to a different interpretation of taphonomic and excavation damage, application of different 'forensic' methods and/or an idiosyncratic incorporation of aspects of non-human primate morphology. Other differences, and in particular upper lip height, are less justifiable in relation to the skeletal evidence.